Select Committee on Democracy and

Digital Technologies

Corrected oral evidence: Democracy and Digital Technologies

Tuesday 28 January 2020

10.25 am


Watch the meeting

Members present: Lord Puttnam (The Chair); Lord German; Lord Harris of Haringey; Lord Holmes of Richmond; Baroness Kidron; Lord Lipsey; Lord Lucas; Baroness McGregor-Smith; Baroness Morris of Yardley; Lord Scriven.

Evidence Session No. 7              Heard in Public              Questions 84 - 98



I: Will Moy, Chief Executive, Full Fact; Ed Humpherson, Director-General of Regulation, Office for Statistics Regulation; Jenni Sargent, Managing Director, First Draft; Allan Leonard, Editor-in-Chief, FactCheckNI.




Examination of Witnesses

Will Moy, Ed Humpherson, Jenni Sargent and Allan Leonard.

Q84            The Chair: Welcome, everybody. Before introducing yourselves, I am required to read out the following, for obvious reasons. As you will know, this session is open to the public. The webcast of the session goes out live and is subsequently accessible via the parliamentary website. A verbatim transcript will be taken of your evidence and put on the parliamentary website. You will have the opportunity to make minor corrections, for the purpose of clarification or accuracy. Perhaps you would introduce yourselves for the record, and we can then begin with the questions.

Will Moy: I am the chief executive of Full Fact. Full Fact is a charity that seeks to promote informed public debate, and is known as the UK’s independent fact-checking charity.

Jenni Sargent: I am the managing director of First Draft. We work internationally and are non-profit. We work to support any information providers that have a role in the information ecosystem to address the challenges of what we refer to as information disorder.

Allan Leonard: I am the editor-in-chief at FactCheckNI, based in Belfast, Northern Ireland. We were established in 2015 and are Northern Ireland’s first and only dedicated fact-checking service.

Ed Humpherson: I am director-general for regulation at the UK Statistics Authority and head of the authority’s Office for Statistics Regulation.

Q85            Baroness McGregor-Smith: Good morning. Could you tell us what effect misinformation or, as I think you just called it, information disorder has had on public debate in the UK? What have we achieved to date that is positive and reduces it, and what are the challenges?

Jenni Sargent: Referring to information disorder reflects how complex this challenge is. We all know that the term “fake news” has become prominent, but it goes nowhere close to reflecting exactly why we are in this situation. There is no question that the rise of the online space has impacted how information is shared and how the emotional drivers of why somebody would choose to share something are being manipulated. The lack of editorial oversight, in the fact that anyone can now become a publisher, means that when it matters most, at a time of heightened uncertainty—whether a democratic moment, an extreme weather event or a health crisis—anybody who wants to disrupt discourse can utilise online spaces to shift the narrative based on how they want people to perceive information. We are trying to raise awareness of the tactics and threats that are emerging.

Ed Humpherson: The statistics authority is here to ensure that statistics serve the public good, and are a reliable and trustworthy evidence base. In a world in which data and information are abundant, people can feel bombarded by information from all sorts of sources. That velocity of information, we think, creates a Gresham’s law risk that the bad data drive out the good, and that people no longer know what they can rely on or believe in. We see our role as leaning against that, but not in an all-purpose way. We are focused up stream, in the sense that, in any exchange of information, there is the producer of the information, who is the person doing the speaking, the intermediary, who provides the space or platform, and then the hearer, who is the user of the information. We focus on government as an upstream provider of information and statistics, providing a reliable, impartial evidence base. There are risks to that impartiality, which it is our job to inhibit, but we see our role in the provision of a reliable evidence base for public use.

Allan Leonard: FactCheckNI was established to improve civic discourse in Northern Ireland. We were established with a grant to encourage active citizenship and participatory democracy. We see our role as twofold: first, to provide to the public free-of-access fact-check articles, to set the record straight; but education and training are also important to us. The idea is not to tell people what to think, but to have people think for themselves. When we write our fact-check articles and do our training, we always emphasise the value of critical thinking. I like to start some of my training sessions by saying, “The best fact-checker in the room is you, because there is no way any single fact-checking organisation can fact-check every claim in the world”. It is imperative to educate as many people as possible about the role of critical thinking, and that is a big part of our role.

Will Moy: Bad information can ruin lives. It damages people’s health. It promotes hate and it hurts democracy. We now see people suffering from curable diseases because they have been misled by false information about vaccines. There is false information about public health issues related, for example, to the rollout of 5G mobile communications technology. We see terrorist attacks sometimes promoted by people who have been radicalised by false information online. We have communities in this country whose experience of life has been really harmed by false information and beliefs about them, created by misinformation and disinformation, not just online but in the media. Of course, it hurts democracy. It hurts people’s faith in and ability to participate in democracy. We did research during the general election, which we would be happy to share with the Committee, showing that significant numbers of people are less likely to vote because they feel unable to trust the information they are hearing.

Not all bad information ruins lives; it is one of the possible outcomes. People being wrong about things on the internet is not a public policy problem that demands a legislative response, but we need to recognise that there is a spectrum of experiences and some of the consequences are very serious. As a country, we should celebrate the many things we do right to tackle this problem. A respected and trusted public service broadcaster, of which we still have more than one, is an important part of informed democracy and healthy public debate. We have freedom of speech; we have institutions such as the House of Commons and the House of Lords Libraries informing our public debate; and we stand up for standards in public life with organisations such as the Office for Statistics Regulation. As a country, we have a lot of building blocks in place that can help tackle the real harms that come from bad information. By and large, we are yet to move them into the world we now live in where, instead of most people using a few sources of information, they receive information from thousands of different sources, and are finding it harder to know what to trust.

Full Fact has a few roles in this. First, we are a direct source of information for the public. We had 10 million users on our website last year. We serve millions more people through partnerships with Sky, ITV, the BBC and other media. We are embedded in Facebook, Google Search and Instagram. We reach millions of people directly. We also challenge people putting bad information into public life. We ask people to correct the record when they need to, whether they are the Prime Minister, a business or a charity. We get people to correct the record, on the record, to stop the spread of specific unsubstantiated claims. We have a track record of doing that: we made 87 correction requests last year, and followed up on our fact checks on more than 120 occasions in total.

Then we use the evidence from thousands of fact checks, from each experience of a piece of bad information being made and spreading, to identify the patterns of behaviour in that information ecosystem. What are the pinch points where changes can make a difference to have a healthier, better informed public debate? It is our job to bring that evidence to decision-makers who can make a difference, whether it is encouraging Facebook, Google or Twitter to change their policies, encouraging the Office for National Statistics to communicate its information more clearly, so it is less liable to be misunderstood, or working with and training journalists. We look for the places where some people’s changes can affect millions of people and improve the quality of public debate.

Baroness McGregor-Smith: We are all citizens. With all this information everywhere, during a political campaign, how would you fact-check a claim? What would you do individually? Allan is talking about the critical thinking that we, as individuals, should apply, but there is a wealth of information on the web. What do you do to check that facts are right? I understand what your organisation does, but what do you do? What would you suggest citizens do to ensure they get a balanced debate?

Ed Humpherson: With my professional hat on, I can give a developed, technical answer, but as a human being, so to speak, my approach is to go the extra click to find out what sits behind a piece of information I encounter. I find that phrase “going the extra click” is a way of reminding myself to have some critical faculties.

Allan Leonard: There is a model we have in one of our projects: think, check, share. As an individual, if I saw a claim, before I retweeted or shared it, I would think. During the last general election, we and Full Fact both participated in a news campaign sponsored by Facebook, which promoted our toolkits. Basically, you go to the page and ask yourself certain questions. Do you know the source of this information? Do you already trust it? Why is this showing in your news feed? Who gains from you reclicking it? You get those questions in your head. Every individual should be thinking about some essential questions.

Jenni Sargent: We predominantly work in the online space and we prioritise digital verification, but we train journalists predominantly because it is unrealistic to expect members of the public to develop these skills. There are many quick checks you can do. You can reverse-search an image to see when it first appeared on the web. Very often, things are taken out of context and, sadly, it is not always as simple as something being true or false, or right or wrong. The reason we are in this predicament is that a lot of the information you see has grains of truth in it. A video may contain a claim that is accurate but presented in a misleading way, so it can be taken out of context. Developing these skills when you are up against sophisticated tactics is quite a big leap for members of the public to take.

To echo the point on motivation, I personally always look at the motivation behind it and whether that comes from a media outlet or a member of the public. If I have received information online, I will check what the motivation may have been, then personally apply verification skills to track the original source.

Will Moy: There are three questions in our toolkit. Where has it come from? What does it really mean and do you have the whole picture, going that extra click? How does it make you feel? You need to be cautious and recognise when something is winding you up. Those are useful guides for me. As you asked the question on a personal level, the starting point is the old maxim that nothing matters very much and few things matter at all. My starting point is to read less news. Trying to be constantly informed is not particularly helpful. There is a challenge with the speed of modern information flow and the temptation that many have to try to keep up with it. You may feel like you are becoming more informed, when the opposite is often happening. Very few things really matter. Trying to identify the things that really matter and spending time on them is important to me.

I think I speak for everyone on the panel in saying that, even as specialists in this field, we can be deceived. There is no magic recipe for spotting misinformation and disinformation instantly. Recognising your own fallibility is part of the picture.

Q86            Lord Lucas: One feature of local digital democracy software is the ability to have a curated set of uncontested information on which people can focus their debate. Is there scope for creating that on national questions? It can be hard for someone who does not have a lot of time and does not know their way around to find where the data is, let alone identify what is reliable.

Will Moy: Yes, there is scope to do that. The House of Commons and the House of Lords Libraries both illustrate that that can be done. The Office for National Statistics, at its best, illustrates that that can be done. They all primarily serve a certain kind of audience. The skill is taking that quality of research and analysis, and the careful, structural and individual attempts at impartiality behind it, and translating it for a general audience. There is a reason that the BBC’s mission is to educate, entertain and inform. Entertainment is part of the picture and, until you have somebody’s attention, you cannot do anything with their attention. There is real skill in good journalism that manages to get people’s attention and then do something useful with it. High-integrity journalism is a really important part of that picture.

Ed Humpherson: I have two points to build on Will’s answer. First, if you look at the range of work that we do regulating government statistics, you will find a drumbeat from us saying, “Do not regard the immediate policy users as the only people you are serving. In fact, you are serving a much bigger public debate, which often requires a more accessible, coherent and linked-up way of presenting information”.

Secondly, I was recently giving evidence before the Treasury Select Committee in the Commons about regional imbalances. I made the point that, if you want to understand regional balances in the UK economy, you have to go to multiple places. You have to look at GDP figures, employment figures, well-being figures and various others. It struck me that a real service could be provided by creating integrated datasets that combine all those things in one place. It is a very well-made point.

Jenni Sargent: Data voids are a real threat right now, inasmuch as people actively search for information when they need it most. This is the area we need to address: when people proactively search for information, what information is returned at the top of those search results? Regardless of where accurate information exists, if it does not hit the first three search results and there are no indicators of credibility to the public, as to why they should understand and believe that information, it is an open goal for anyone looking to drive a different narrative.

Allan Leonard: In Northern Ireland, there was an instance when a journalist had used data from NISRA and elsewhere on migration. The data was correct, but not in the right context or applied in the best way. Siobhan Carey is the chief executive and general registrar of NISRA. She and her team identified that part of their role is to make all the data that they collect usable. That is important, as their statutory obligation is not only to collect data, but to make it practically usable to the public they serve. They created an amalgamation of relevant data products, called NI: IN PROFILE. This is to engage with ordinary journalists so that, when they are covering a particular topic, they have to hand a set of the most relevant data that they should be using. There is a lot of scope for this, because journalists are an intermediary to a broader public. I am exploring extending this within Northern Ireland.

Ed Humpherson: We really like NI: IN PROFILE. It is a great product.

The Chair: In the interests of transparency, could all Members of the Committee register any interests, because it is our first meeting since the election? Any interests that you might bring to the table have to be recorded.

Q87            Baroness Morris of Yardley: Thank you for the reminder. I register my interest as a chair of and adviser to the Birmingham Education Partnership and a trustee and adviser of the Institute for Effective Education, which is financed by the Bowland Charitable Trust.

To some extent, this question follows from the answers you have given about how you would respond and try to check facts. It is about the expectation we have of the public, and the knowledge and skills they should have in spotting and not spreading misinformation. You have already said how difficult it is. When I listened to what you do, I wondered whether we can really expect every member of the public to develop those skills and to do that on every occasion that they come into contact with data. Is that expectation too high? What should our expectations be of the skills the public have? I suppose the answer has to be education, but can that meaningfully reduce the spread of information or does identifying misinformation require an amount of expertise or time commitment that is not to be expected from the general public?

Jenni Sargent: Controversially, I completely agree with you. There is a big push in our industry, at the moment, on media literacy and this idea that it will be the saving grace: we just have to integrate it into school curricula and we will be fine. Consider the sophisticated tactics at play: people are being deliberately targeted with advanced techniques. Tactics of persuasion, historically, are driven by many different skills. When consuming news, most people have an element of digital literacy and understanding of motivation. But there are so many factors at play that, from my point of view, it is definitely the responsibility of the platforms and publishers to provide indicators of credibility. The consumers of information cannot and should not be expected to perform these additional checks on everything they see. It does not feel realistic and it is not the way that we normally communicate.

We have none of the verbal or visual cues that normally accompany communication. The reason that we have emojis now is to help people understand sentiments when messaging online. Most pieces of information, when presented on Facebook and other online news forums, have none of the indicators that we might recognise, such as newsprint font or the paper colour of the Financial Times. We have no indicators to suggest why something on Facebook is presented differently from another thing. We can and certainly should inform on these issues from school age but, as we start to discuss the threats and challenges of misinformation, it is a giant leap to think that media literacy will solve the problem.

Allan Leonard: I advocate media literacy. I have had this discussion elsewhere on the role of technology platforms versus education of the public. It is a supply and demand issue. I do not disagree that there is no silver bullet that will resolve this. A lot comes down to trust, which is a difficult thing to achieve and standardise. I just make a distinction between claims that are made in public and microtargeting. My view is that, for claims made in public and broadcast publicly, media literacy, general awareness and fact-checking organisations can respond very promptly. Resolving the issue of microtargeting political adverts and any other form of advertising is a bigger challenge, and education on its own is not going to resolve it. There needs to be much more transparency about who is targeting whom, why and who is to benefit. That is where software solutions could play a role.

Will Moy: Full Fact, as a charity, is ultimately about citizenship. In a democracy, we are all members of the general public. We are all the most important people in a democracy. We have to believe in the role of the individual citizen as an important part of having a healthy democracy. Our skills and knowledge and the information we have are a central part of that. Yes, I think education is part of the response to this challenge, but we have to think about what the goal is. If the goal is to eliminate belief in things that are not true, first, that would be a remarkable thing for a Government to set their mind on and does not stand up to much scrutiny. It is obviously dangerous and impossible to go down that road.

I think the goal is to reduce harm from false and misleading information by taking proportionate steps, in the same way that any form of regulation tries to be a proportionate response to clearly identified harms. In that sense, some information matters more than others. Before you share something online that purports to be a cure for cancer, you might start by thinking that you have a higher responsibility than when you share something that purports to be a video of a cat jumping over the Empire State Building. People are capable of making these distinctions. We need to think about the frictionless environment for sharing information and ask people to think more clearly, as the thing they choose to share might not be equivalent to something else that has a real impact on people’s health, votes, financial choices or lives.

The case against education is easy to make. If you educate schoolkids, it will take 50 years to filter through to the whole population and make a difference. The generation most likely to vote is 50 years away from school, not five years. Education is not perfectly recalled. We cannot educate the adult population in any way that requires sitting down for a long while and gaining new skills. There is no scalable way of doing that. It is easy to point out the difficulties and the reasons why education is not a sufficient response to this problem. The challenge is to find the one sentence of advice that we could tell everybody to think about. What is the equivalent of “stop, look and listen” that has a proper research base behind it, which might be a useful first step for everybody? It might be “take the next click”, for example. That could be useful. We have evidence that the right educational interventions can make a difference to people’s choices, and that it can be a sustained difference.

Ofcom is running the Making Sense of Media Advisory Panel at the moment—I am a member. It is a civil society group trying to co-ordinate the many efforts being made on media literacy at the moment. We are very pleased Ofcom has done that. That co-ordination is necessary. It now has an evaluation subgroup, specifically focused on understanding what is effective in this area. That is a good place to start. It is notable that the Department for Education is not actively involved in that effort, at the moment. I recognise the many pressures on the curriculum and the trade-offs of putting anything in it or taking things out. Teachers are already trying to improve critical thinking; it is not like they have not thought about this, but the Department for Education ought to be actively involved in those conversations and thinking about its role.

Baroness McGregor-Smith: I am a non-executive director at the Department for Education, so will feed that back. It is something that we will definitely pick up.

Q88            The Chair: I feel like a cracked record in saying this but, 17 years ago, Ofcom was given statutory responsibility for Digital Literacy, and yet here we we are. How is that possible?

Will Moy: There are two reasons. First, Ofcom has many statutory responsibilities and I am not sure it saw this as central to its role. The information environment 17 years ago was essentially about understanding your way around five TV channels and 10 newspapers. Ofcom has increasingly recognised that media literacy is a strategic necessity for the future. It has moved from being a part of Ofcom’s role to something that needs to be thought of as more central. I cannot speak for Ofcom, but that is my impression of the way its thinking is developing. The reason for that is that media literacy, whatever we mean by it, is harder than it used to be, because the information environment around us has changed.

Ed Humpherson: This is really not my area of focus. As I said, I focus up stream on the provision, creation and dissemination of statistics. But I think of this as a value chain, with a producer, intermediary and hearer. To focus on digital literacy as mitigation puts a lot of onus on just the hearer, the person receiving. It is important that people have the confidence and the mnemonics or rules of thumb, as Allan was mentioning, but you can only be literate with something readable. Literacy without something readable is meaningless, which is why you have to think about this holistically. It is not simply about equipping people with digital literacy, but thinking about what intermediaries do to make the information provided checkable and verifiable, with digital imprints, and what providers do to make information accessible and meaningful.

Baroness Morris of Yardley: I tend to be where Jenni is, because I think this is different from anything else we have tried to teach, and we have not come to terms with that and worked it out yet. Is there a problem in our doing this, when we have not actually decided what we are trying to do? I can see how difficult it is, and my head is there as well. We have not decided whether we can educate everyone to fulfil that role of active, discerning citizen, or whether we have to put something else in place. If it is the alternative and we have to trust the intermediary organisations, how confident are you that people in your sector are willing to be checked, verified and have that little tick that says the public can trust them?

Jenni Sargent: It is an interesting point, because we predominantly work with journalists at the moment, and our goal is to bring them together to agree transparent indicators that will help their audiences understand why their work can be trusted. That could be as simple as the transparent process that demonstrates the principles that underpin good journalism. This merging that we have now, in which you cannot believe anything and journalism itself is potentially being undermined, is a real issue. My fact-checking colleagues do extraordinary work, but I worry that fact-checking is becoming an industry in its own right. There are potential threats to the value of journalism and understanding good journalism.

We have an initiative called CrossCheck, which is built on this idea that newsrooms endorse and verify each other’s work. These logos go side by side to demonstrate to the public that not just one outlet believes something, but it has been agreed and verified by a number of outlets and, potentially, other expert organisations. The World Health Organization’s logo, for example, might appear to show that it also endorses this. My frustration, when it comes to where the onus of responsibility lies, particularly for the topics that Will highlighted, is that this should be ingrained in the production of the material.

We recognise a copyright symbol, for example, and can begin to understand what that means. I do not think a symbol, such as a tick, is enough in its own right, but it can be instinctive, so that, whenever you see that and click on it, it displays the information in more detail and the producers of information are compelled to show their work in that way. Then we come a step forward. Instead of asking people to do the checks themselves, they can simply have quick access to the information they need.

Q89            Lord Lucas: In a way, that has answered the question I was going to ask. If you are looking up—and I just flipped into this—the suicide rate among transgender people, the first thing you get is Stonewall with a completely rubbish statistic. Nothing in the rest of what Google is offering gives you any opportunity to check that. How many members of the public know statistics anyway? Half the stuff you want is behind a paywall, so it has to be made easier for people, as Ed Humpherson says, to do one click onwards, but not 101 clicks onwards. Would the big media companies support that?

Jenni Sargent: The idea of being accountable to each other is something I would like to encourage.

Allan Leonard: Will referred to the BBC as a public broadcaster and its duty to inform. FactCheckNI participated in the BBC’s Young Reporter programme in Derry/Londonderry a couple of years ago. That is a good example of how the BBC can work with fact-checking organisations and others to educate children about trustworthiness and transparency. We also participated in the FactCheckEU programme during the European elections, on the back of the CrossCheck programme. FactCheckNI and other fact-checking organisations worked with media organisations to scrutinise claims that were being made across Europe.

I think it can be done. I am pausing, because I think this Committee should consider who the catalysts and champions are. Is it the BBC or Ofcom? Who are the conduits to bring together evidence and research-proven effective measures?

Q90            Lord Holmes of Richmond: I declare my interest as a member of the board at Channel Four Television. What are the characteristics of information that achieves public trust? How can we ensure that the public trust and accept high-quality information, such as national statistics, and can differentiate between legitimate and illegitimate sources of information?

Ed Humpherson: It is always a mistake to regard statistics as just numbers. They are numbers with a social life; they exist in a context. We therefore embedded into our code of practice, with which all government departments must comply, three concepts. First, they must be trustworthy. That is not about the statistics themselves, but the signals that the organisation gives about the impartiality with which it has compiled its statistics. You will be familiar with the concept of trustworthiness, because you heard evidence from Baroness O’Neill, who was the progenitor of the thinking on this. Then there is quality, which is about what the data are, where they are from, how they are collected and the inherent biases that must be recognised in them. Most important, in a sense, is value, which is what questions this information is available to answer.

Those three pillars, trustworthiness, quality and value, underpin our thinking about everything. They are about statistics being much more than numbers. They are numbers produced by people in organisations and provided to people to understand the world. They are numbers with a social life. Recognising that and not cleaving statistics out of their context, regarding them just as clickbait numbers or the big numbers to win the argument, is the key to creating a trustworthy environment.

Allan Leonard: Full Fact and FactCheckNI are both verified signatories of the code of principles set forth by the International Fact-Checking Network at the Poynter Institute. I can summarise this in one sentence: trustworthiness is reflected in our commitment to non-partisanship and fairness, to transparency of sources of information, to transparency of funding and our organisational details, so how we are run. A lot of those are qualitative, but we are independently assessed every year and have to show evidence that we have satisfied each of those criteria. To answer the question about trust, trust is earned, but you need to provide sufficient evidence and the evidence you produce may not be sufficient to the public. That is life.

Will Moy: Can I throw a question back to you? Imagine you are taken on an away day. I doubt they do that for House of Lords Select Committees, but you are taken on an away day and asked to do the trust game of falling back into one another’s arms. You do this game 100 times, and 99 times you successfully catch the fellow Member of your Committee but, on the 100th time, you drop them. Will they ever trust you to do that game again?

Trust is not a single thing; it is a collected set of experiences. It is the sum total of your experiences of an organisation, a source or a piece of information. It is, as people say, easily lost. We often forget in this area that Government are losing that trust regularly, by saying things that are not trustworthy, giving that trust away and sacrificing it. The media do that regularly, as do many other powerful actors in public debate. People remember, so people’s default and rational response to this situation is to be generally distrustful. Only about one in five of us generally trusts politicians to tell the truth. Only about one in five generally trusts journalists to tell the truth. That does not mean we think they lie all the time; we think they lie enough that it is not sensible to trust them. Similarly, if you are a non-executive member of a board who is misled by the people who report to you, you would quickly become distrustful of them.

My second concern in this area is about the person who was must trusted to give information during the EU referendum. I do not know if you know who that was but, according to a poll in the October before the referendum, the answer was Martin Lewis, the money-saving expert. Martin Lewis, to be honest, knows nothing about Britain’s relationship with the EU. He is an extremely knowledgeable and thoughtful man, but that is not his area of expertise. He is, however, easily identifiable and well known to people. People have a consistent record of trusting him to give them advice about something that has been important to their lives for years, their finances, and he is clearly on their side.

If I start looking at public debate for the people and institutions that are identifiable, knowledgeable and clearly on your side, I come up with a very short list indeed. We need to start thinking about which public institutions can be trusted in that way and how to make them stand out. There are those such as the Institute for Fiscal Studies at one end of the spectrum and public service broadcasters at the other end. We need to recognise how valuable those voices are in our public debate, protect them and build on the extraordinary assets that we have in the UK, which some other countries do not.

My third vignette is about Lidl, the supermarket. As you may know, not everybody believes that Lidl sells the best-quality food. They get a certain amount of grief about it online. Lidl did something interesting; it did a series of adverts called “Lidl surprises”. You can find them on YouTube. They put on national TV tweets that people had sent saying, “This food must be rubbish”; then they said, “Okay, come and see it”. There is a lovely ad, where a woman who has tweeted critically of Lidl’s steaks is taken to Scotland to see the farm, to meet the cow, to meet the farmer and to try the steak; I hope it is not from the same cow that day, but I do not know. At the end, she says, “I am sorry; I was unfair to you”. The whole story is shown from beginning to end, so that people can make up their own minds about it.

There was a time, perhaps, although I do not believe it, when people would just defer to authority and take its word for things, but we all know that that time is over. We have not adjusted, particularly in the public sector, to the need to tell the whole story, so people can make up their own minds, from those original sources that Ed talked about. There are reassuringly boring spreadsheets, with reassuring normal statisticians looking at them, with no hidden agenda. Tell that whole story, so that people can see it, make up their own minds about it and trust it, in a two-way conversation, not handed down from an old-fashioned source of authority. Those are my thoughts on trust. First, it is easily broken. Secondly, we need identifiable expert people, who are on our side. Thirdly, we need to tell the whole story, so that people can make up their own minds about what to trust.

Jenni Sargent: I am a bit concerned about the perceived elitism aspect, when we discuss this. There is a challenge here. I believe people trust things when they feel safe and comfortable. There is an element of tribalism to some of the online debate that we witness, which is the sense of deferring to information that confirms your own beliefs, feeling safe and trusted in that environment. We are seeing a rise in the spreading of misinformation within closed trusted groups, on closed messaging apps, such as WhatsApp. These smaller trusted groups of peers and family are what we mean by trust. I completely agree that institutions driving accurate information are essential but, at the same time, we need to understand and bear in mind the narratives at play in why people might choose to believe something.

Ed Humpherson: I mentioned earlier the importance of going the extra click. To indulge in some real-time fact-checking, Will was quoting figures about whom people trust to tell the truth, and I think he was quoting from the Ipsos MORI Veracity Index. We should not despair, because there are some professions that people increasingly trust to tell the truth, including scientists and civil servants. People have consistently increased the rate at which they trust civil servants to tell the truth from 40% in 1983 to over 65% in the most recent survey. I think that is an interesting observation. Giving signals of trustworthiness sometimes works.

Q91            Lord Scriven: Jenni took the words out of my mouth. I was sat here thinking, “We are having a very sterile debate about institutions”. Look at the statistics of ethnicity, age, social class, et cetera; some of the institutions that you are talking about just do not relate to people’s real lives. First, do you have any evidence that fact-checking and what you do changes people’s minds from irrationality to a more rational view about what is true? If so, what is the breakdown by gender, ethnicity, et cetera? Secondly, beyond that, to use Will’s words, we have to get into the world we live in now. When given a fact, a lot of people completely refuse to believe it, because of the stuff that Jenni has just said, where they are going to a closed WhatsApp. Is there anything above what you do that can augment that deeply held view as, no matter how many facts you give them, they will not believe it? We talked about issues of citizenship, but is there anything that particularly needs to be done in the online digital world? It comes back to Ed’s point that it is not just the consumer of information. This is quite key, because there is a balancing act at play between information-givers and receivers.

Will Moy: That is exactly the right challenge. Last year or the year before, Full Fact, Africa Check and Chequeado issued a joint statement called, “Fact checking doesn’t work (the way you think it does)”. If you believe that fact-checking can work by simply giving people information, and people reading those fact checks and then changing their minds, you are on a hiding to nothing. Most of our work and the change we are trying to achieve is up stream. It is asking people to correct the record and stop the spread of specific unsubstantiated claims. It is looking at the causes of bad information and trying to tackle them. Yes, 10 million users came to our website last year and, yes, we know that it served some of them very well. We had a phone call from one woman who said that she had gone online to try to work out whether to vaccinate her children, and she could not find any information she trusted. Then she found our work, she felt able to trust it and there are now two children who are vaccinated, who would not have been without the information she found from Full Fact. We know it can make a difference to people.

We also know, as you say, that people do not just process information; they process it in a social context. We all process it through our psychological biases and our emotional predispositions. If you wanted to change people’s minds at scale, you would never choose fact-checking as the way to do it. You would do a biased campaign that targets people’s emotions. If you want to change people’s attitudes to NHS spending, immigration, the right balance of public spending and taxation or whatever, pick a fight and join in. That is what a democracy is. If you want to convince people of things, you should use facts, which are part of convincing people, but you should also use emotions and all the techniques of campaigning and communications.

Fact-checkers are not the answer to how to get everybody to believe things that are true. We are part of the answer to how you give people the opportunity to make up their own minds about what is true and what is not, and to how you challenge people in positions of power and responsibility, when they choose to mislead others or when they accidentally mislead others. The idea that there is anybody who could sit in some central position, bestow accurate information on the world and be gratefully thanked for it is borderline silly, to be honest. It surprises me how many people look at fact-checking through that lens, because it is not a viable interpretation. The accountability function is a crucial part of what we do.

Q92            Baroness Kidron: I apologise; I have a short question but a long list of interests to read out. I am a commissioner on the UN Broadband Commission for Sustainable Development, a member of the technical board of the WePROTECT global initiative, a member of the Council on Extended Intelligence, a member of the UNICEF artificial intelligence and child rights policy guidance group, chair of 5Rights Foundation and I speak on digital technologies in childhood.

Lord Scriven: For the record, I am a vice-president of the Local Government Association.

Baroness Kidron: I am sorry for that interruption of service. I am very interested in this idea of the value chain. There is the producer of the information, the intermediary and then—I very much like your word—the hearer, rather than user, at the other end. I was equally struck that all your institutions have a set of values or principles. What set of principles could the intermediaries adhere to that would help to drive a healthier relationship between verified producers and hearers? That is the bit we are getting least action on in this conversation, yet we know they are a powerful force in how we receive information.

Allan Leonard: I will answer that from a slightly different angle. Fact-checking can be described as explanatory journalism. The ambition is to put claims and facts into a context. A lot of daily journalism is just getting out the daily story and narrative. We are based in Northern Ireland, and there is some debate about the role of the broadcast media during and post the Troubles. I attended a conference in Belfast recently and what came across from listening to journalists from South Africa, in regards to their principles and values post apartheid, was duty of care and harm. They revised their editorial guidelines so that, in where they send their reporters, their story selection and how they cover stories, they have a duty of care and a duty to do no harm. As you mentioned, Lady Kidron, we at this table all have that. I would love to see more media outlets pay more attention to their duty of care and to do no harm in the stories they cover.

Jenni Sargent: This is quite nuanced. Facebook and other platforms have directed the behaviour of some news outlets, because of the financial incentives and the way the information landscape has shifted. We all know the term “clickbait”. It is potentially not appropriate to use that to refer to all news organisations, but there is definitely a drive to become more competitive in a smaller online space. How do you drive visitors and viewers to your website, when you are in an aggressively competitive space? The platforms themselves need to provide a better way for information providers that demonstrate credibility not to reduce themselves to these tactics to drive users and viewers, which is part of what muddies the waters.

The financial incentives across the board are the real issues here, because you have everyday users online in most of these free spaces, but the paid-for advertising means you can leapfrog any of these tactics and put out your messages and narratives, sometimes without the transparency needed. I am always shocked by how the platforms do not refer to the news industry more frequently, to learn from what it has gone through in terms of how to report, such as whether to show the moment of death. What decisions have been taken, through years of experience in the newsrooms, that are not applied to online spaces now, because they refuse to consider themselves publishers? I would like them to consider themselves publishers.

Q93            The Chair: Mr Humpherson can I somewhat reverse that question to you, because you run an immensely trusted organisation? What keeps you awake at night? What could seriously damage the credibility of your organisation? How could you find yourself emerging as part of the problem, instead of part of the answer?

Ed Humpherson: What keeps me awake at night is that so many official statistics have the potential to be valuable assets, not just for the policy elites, but for the public more broadly. But that potential is unrealised, because they are presented and communicated in a dry and almost mechanical way: we did this survey and got these results. They are not presented in a way that engages and links into the things that people are concerned about. I worry much more about that than I do about the more pervasive concerns people have about misinformation. I worry about the good information not getting out there.

Will Moy: I have a point of information and then will answer Baroness Kidron’s point. Only about half of people generally trust that statistics on employment and immigration are produced without political interference. Going to Lord Scriven’s argument, there is a clear demographic skew in where that trust lies, so we need to be very careful. The Office for Statistics Regulation is, rightly, a very trusted organisation; as are the ONS and the statistics authority. We need to be careful about where the limits of that trust are, in practice.

Baroness Kidron’s point is a huge question. The starting principle that these organisations should embody, in a free society, is free speech. We need to be careful that proportionate responses to real harms do not crowd out free speech. The idea that those companies are best placed to work out where that balance lies, in all the different countries around the world in which they operate, is laughable. They are doing so because Parliament has failed to act. Instead of having an open, democratic, transparent process defining where that balance lies in the UK, we have Mark Zuckerberg, Jack Dorsey and the rest of them. I do not think that is acceptable, but the responsibility for it is yours.

These internet companies should provide proportionate responses to real harms. They should be actively monitoring harmful content on their platforms, and responding. I do not think they are capable of defining where that line is without democratic oversight. What is a harm that qualifies for intervention and the suppression of somebody’s speech and what is a harm that does not? That is a decision that I would like to see taken with democratic accountability, not just in a corporate situation. To their credit, they have also called for that, to some extent, and it is about time we acted.

Q94            Lord German: I declare my interest as a treasurer of the Liberal Democrats. Can I try to tease out your answer to whether what is happening is credible and your role within it? You already mentioned your connection, particularly Will’s, to the technology companies. One might assume that the public feel that you, Ed, are associated with the Government. That might be the way it appears for a statutory body or whatever. Outline for us the credibility advantages and disadvantages to being connected either to government or to a technology company, in any way whatever. If you have any solutions to the issue of weaknesses and how misinformation flows through, they would be helpful.

Will Moy: Let me start by declaring an interest on behalf of Full Fact. Full Fact receives funding from Facebook. We have grant funding from Facebook for work we are doing to promote technology for fact-checkers internationally and AI in this area. We have funding because we fact-check content on their platform, at their request, to help them identify and respond to false information. We have funding from Google. We were winners of the Google AI for social good impact challenge, which was a global challenge for artificial intelligence in this area, with entrants from more than 100 countries. We were one of 20 winners. We do not take government funding, but we have previously worked with the UK Statistics Authority and the Economic and Social Research Council, both of which are independent from government, but ultimately use taxpayers’ money. So you are aware of the background here, all our funding is declared on our website and available for the Committee, or anyone else, to inspect.

As an organisation that seeks to promote accountability in a complex and contested field, such as public debate, we will be rightly scrutinised. We have to think about every partnership and interaction very carefully. When we work with a major media organisation, we fact-check it. Sometimes, we have to ask media partners of ours to correct the record, including in the heat of a general election. When we work with internet companies, we also come here and tell you what is wrong with their work. When we chose to work with Facebook, our condition for doing so was that we would publish a regular transparency report on its third-party fact-checking programme, saying how it works and how it needs to improve. We published the first one last July, after six months; we are drafting the second one now. That was our way of maintaining our independence visibly and making sure our participation in that programme helped us drive improvement, rather than just be part of it.

We also work with the Government. We work with the Government Communication Service, which is at the front line of tackling disinformation and particularly state-sponsored disinformation, but the Government Communication Service is also one of the major providers of disinformation in this country, as the voice of its political masters. All our relationships are complex and they all involve having to maintain integrity in a place where people could reasonably doubt what is going on and what that relationship is. All we can do is behave to the best of our ability, have good governance in place—we have a cross-party board of trustees, which scrutinises all our work—and hope our actions and our efforts to explain them to the public are sufficient for people to place trust in us. But it is a complex field and it is difficult to get those judgments right.

Lord German: I am sorry for interrupting before we have heard from everybody else, but you have not answered that one question, which is whether it makes you more or less credible to the listener or hearer.

Will Moy: It makes us more credible, in providing evidence about how the internet companies behave and how that can be improved, that we have worked with them, are known as expert by them and have information about what spreads on some of those platforms, because of our work with them. It makes us more credible when talking to, for example, public information bodies about how to communicate effectively in the online world, that we work with the internet companies. It makes us more credible to the internet companies, when we say to them, “We want to publish a transparency report on your programmes”, that we work with government and understand the policy environment that affects them. It makes us more credible to media organisations that we work with the internet companies, and vice versa. There is a chain of credibility in that somewhere, because we work with the top-tier organisations in all these fields.

We work with the House of Commons Library, the ONS, the public information bodies, top universities, top internet companies, the top broadcasters and media outlets. We have something to offer in all those areas. The problem is that each of those organisations is suspicious of the other organisations in that chain, so they also have to see from us something that demonstrates our independence from every organisation we work with. Ultimately, you would have to ask others how we are regarded, but the evidence so far is that all those organisations are willing to trust us and, to some extent, willing to trust their reputation in partnership with us.

The Chair: Ms Sargent, do you have anything to add that is specific to your organisation?

Jenni Sargent: We are also funded by the platforms. We recognise that anyone within the information ecosystem needs to be part of this discussion and that collaboration is the only way we can bring in all those stakeholders. One of our biggest complex challenges, because we work internationally, is that no two Governments are the same; no two news organisations are the same; no two media landscapes are comparable, in terms of who we might consider are credible contributors to this discussion. We are continuously trying to evolve the principles and guidelines around this.

Allan Leonard: Like Will explained, FactCheckNI is also a partner organisation with Facebook’s third-party fact-checking programme. We have similar provisos. FactCheckNI is also part of a European Commission Horizon 2020 research project called Co-inform. We are among a partnership of universities and research institutions. We are there as practitioners. The credibility comes from research efforts that are exploring this and getting input from those of us on the ground doing the work. We can feed back from our experiences, but also draw from the global fact-checking community and what has been shown to work elsewhere.

Ed Humpherson: I am glad this question came up, because I was hearing everybody else declare interests and feeling left out, as I did not have one. We are a taxpayer-funded body. It is an enormous privilege to be a statutory body and have an Act behind us. It confers on us a real responsibility to show that we act in the interests of the public. One way in which we need to be very judicious is in not politicising our interventions. That is not to say we do not make statements that are uncomfortable for one or other of the political actors or parties. We do that, but we are parsimonious in doing it, so that anybody looking at us sees an even-handedness in how we form our judgments.

Q95            The Chair: There are two final questions, and I’m afraid we are running out of time. I have to an interest to declare and onewhich is quite specific. I sit on the advisory board of Accenture Ireland, which makes this question rather more pointed. How well does Facebook’s partnership with fact-checking organisations work? Should it be emulated by other technology companies, and what can be learned from approaches taken by other companies?

Will Moy: Full Fact publishes transparency reports on the third-party fact-checking programme. The last one included 10 recommendations for Facebook about how the programme should improve. Our basic view is that the third-party fact-checking programme is a valuable measure, which is particularly effective and useful in certain areas where clear-cut action is called for, such as tackling election interference, tackling false information in the aftermath of emergencies, including terrorist attacks, and tackling health misinformation. We think it should be adopted by other internet platforms. In our last report, we recommended that Facebook extend it to Instagram and it has done so, which we are pleased to see. We think other internet companies should follow suit.

We raised two thematic challenges for Facebook. It needs to be more transparent about the programme and how well it works, and to allow that to be assessed independently. We also need to talk about how we scale it, because this is an internet-scale challenge. That is our broad view of the programme. As I said, we are currently drafting our second transparency report on it.

Jenni Sargent: I have some concerns about it, actually. It has come a long way since it was first evolved at the beginning of 2015. There is no question that it has value. I have concerns about whether it is scalable. There are real issues about content moderation, which is being outsourced by the platforms. I cannot see how this does not fall into the expansion of what they refer to as “third-party”, but it is a direct financial relationship. Where Facebook pays organisations, it raises questions around incentives and how to ensure that there is no financial incentive to publish a fact check. Even though IFCN, as Allan referred to, has some excellent principles and is thoughtful about how to address these challenges, I worry that fact-checking organisations are sprouting up in order to receive this financial compensation. There is a risk of erosion of trust, if all we do is publish what is not true. I would like to make sure there is a similar stream about accurate reporting and ensuring we are raising awareness of what is true and not just what is not.

Allan Leonard: That is one of our major concerns. Generally, FactCheckNI is content with its participation. The volume of our work is smaller, because of economies of scale and we represent a smaller population. Positively, the participants are verified signatories of the IFCN code, which is good, as an expansion would have to go through that third-party process.

As the previous two speakers have said, the challenge is that Facebook is global. A one-size-fits-all approach is not going to work in a variety of cultural, operational and legislative frameworks. We all need to remember that. Also, there is a reputational risk in the sense that Facebook will not apply treatment or financial payment for anything it deems political. That is fine, in the sense that that is its policy. You have to challenge it on its policy. The reputational risk is that the wider public can misassociate a fact-checking organisation as not doing political fact-checking, and that is just not true; the majority of our fact-check work is political speech. It is just that that speech is ineligible to be considered, so that is an identifiable risk.

Likewise, on bias, I am very concerned that, in trying to improve algorithms and learning, if you are feeding the algorithm only evidence of claims that have been proven false or partly false, it does little to educate anyone about the effectiveness of fact-checking. When FactCheckNI researches any claim, we publish the verdict regardless of what it is, whether it is accurate, accurate with consideration, inaccurate with consideration, inaccurate or unsubstantiated. We participate in the programme, but I just highlight those issues.

Ed Humpherson: I would say what they said.

Q96            Lord Lucas: If the Government could do any one thing to ensure better informed public debate in the digital age, what should it be?

Will Moy: It is the wrong question. I say that with great respect, but great seriousness as well. There is a tendency to see this issue as a problem for which we are looking for a solution. It is actually a new area of policy, which we will be debating for at least the next 30 years, in which we will end up with a different balance between free speech and harms, with different kinds of harms being considered, than we have now. I have no idea what the outcome of that discussion will be, but it is profoundly important and needs to happen in the democratic context. The Government are making the mistake of thinking that they can create some kind of statutory duty of care, then outsource all those profound choices to a regulator, when they are democratic choices that need to be democratically debated. There is not a solution here, nor one step to be found. That is my push-back to you.

More broadly, we are missing the trick, which this Committee has brought out. There is something called the Swiss cheese model of accident prevention, which is the idea that everything has holes in it, but you cannot find your way through a block of Swiss cheese, because the holes do not line up. When a major accident happens in transport safety, people look for a chain of things that have gone wrong for the catastrophe to have occurred. The opposite is true in this field: being well-informed is a remarkably surprising thing. You need somebody to have asked smart questions, good research and evidence-gathering to have happened, that to have been communicated clearly and reached someone who can reach a large audience, and the audiences to have taken it in. Every one of those steps is surprising. We need to think in a more serious way about that ecosystem and how we can promote informed debate at every step, and to stop looking for one answer to the problem.

Jenni Sargent: I agree. I would like to see some action. Debate is great, but 30 years of debate would put us in a dangerous place. There may be an equivalent of COBRA for certain moments. We now refer to the internet as a hostile environment, so journalists have hostile environment training and we now say that anyone working online should receive the same kind of support. Harassment and harm tip into disinformation, threats and challenges. How that can easily escalate offline to moments when people are scared or fearful is a real issue. Rather than trying to tackle the whole problem, we could look at the definition of harm, as Will alluded to earlier, and the moments we are really concerned about that have a greater impact, through misinformation spiralling and escalating out of control, which is possible.

Will Moy: To clarify, based on what Jenni said, my proposal is that you legislate little and often, as we understand parts of this puzzle. Be prepared to update that legislation as our understanding grows, instead of trying to legislate once and for all.

Allan Leonard: I agree with much of what Will said, to be honest. As I mentioned previously, it is about identifying well-trusted organisations and thinking of synergistic ways for them to engage with fact-checking organisations, universities and research to, as Will explained, help promote the pathway of good information.

The Chair: Mr Humpherson, the question to you, as a statutory body, is what could we suggest that might help you?

Ed Humpherson: I was going to offer advice on the one thing that government could do. Do not simply look outwards and negatively. In my value chain, by all means educate and support the hearers and regulate the intermediaries, but also think about the Government as a provider of information. Spend as much time thinking about what it means to inform as it means to misinform.

Lord Lucas: Actually, Will, democratic pressure is working already. If you google, “Do vaccines cause autism?” it says, “No, no, no”, in capital letters. If you ask, “Is 5G dangerous?” it is about 50:50. If you ask, “Is climate change real?” it is clearly not happening.

Q97            Lord Harris of Haringey: I go back to the point Ed made earlier. I declare my interests, because I have not spoken yet. I am a board member of Cyber Security Challenge, chair of the Independent Reference Group of the National Crime Agency, and chair of National Trading Standards, none of which is to do with my question. You said earlier that you are quite parsimonious about criticising political parties, and that you try to achieve balance. If—and I appreciate this is a rather large “if”—one party consistently misused statistics and another party never misused statistics, how would you achieve balance?

Ed Humpherson: We do not have the objective of achieving balance. If I conveyed that impression, that was wrong. If you look at what we do, you will see that there is a balance, but we do not have a sense that, if we have criticised one party, we have to find a criticism of another. In a way, to use the football analogy, we do not even look at the player; we look at the ball. What are the statistics in question? Is there a risk that the statistical evidence base is being misinterpreted or misrepresented? If it is, we will seek to clarify it.

Q98            Baroness McGregor-Smith: In all our debate, no one has talked about how we should maybe say to citizens, “Come off all this stuff”. I say that to my kids. They are off a lot of things. Protect your privacy and yourself as the individual. There are things you can do to protect yourself. You do not need to be on everything. To this point about social media feeds where you can be attacked, there is a very simple answer: do not be on them. You would not open yourself up to attack walking down the road in the middle of the night. In the same way, why would you do that on the web? Why would young people listen to all that? It is just an interesting point that has not come up, but I think a piece of work needs to be done to explain, particularly to young people, that you do not need this in your lives. We survived without a lot of it. There are things you can do to support the individual, particularly against mental harm.

Jenni Sargent: The opportunity to be more politically informed, be more politically engaged and receive more accurate information, because of the internet, should not be overlooked. My concern is that the less people use these platforms, the less familiar they will be with how they might be manipulated and receive information. Regular users are more likely to understand the etiquette and challenges.

Baroness McGregor-Smith: My challenge back on that for young people particularly is that, certainly my kids, they are coming off in droves. They want to protect themselves from what can be said on these platforms. Yes, they are pretty tech-savvy, but that is a valid thing to do.

The Chair: If you have any thoughts, please feel free to write to Baroness McGregor-Smith. Thank you very much indeed for your time.I am sorry we have kept you.