Written Evidence to Science, Innovation & Technology Select Committee
Social media, misinformation and harmful algorithms
March 2025
- About ISBA
1.1. ISBA represents UK brand advertisers. We are the only body in the UK that enables advertisers to understand their industry and shape its future, because we bring together a powerful network of marketers with common interests, empower decision-making with knowledge and insight, and give a single voice to advocacy for the improvement of the industry.
1.2. ISBA is a member of the Advertising Association, and represents advertisers on the Committee of Advertising Practice (CAP) and the Broadcast Committee of Advertising Practice (BCAP) – the sister organisations of the Advertising Standards Authority (ASA) which are responsible for writing the Advertising Codes. We are also members of the World Federation of Advertisers (WFA). We are able to use our leadership role in such bodies to set and promote high industry standards, as well as a robust self-regulatory regime.
- Response to Questions
To what extent do the business models of social media companies, search engines and others encourage the spread of harmful content, and contribute to wider social harms?
2.1. As the trade body which represents brands advertising in the United Kingdom, ISBA welcomes the Science, Innovation and Technology Committee’s interest in this subject, and the fact that it was the first inquiry launched once the Committee was constituted after the General Election. With social media platforms now so integral to the media lives of the population, and with the particular focus on safety – especially among children and young people – it is right that there is accountability to Parliament.
2.2. Social media platforms currently operate on the basis of a bargain: the platform is free, in exchange for the data of the user. This data is a powerful commodity for advertisers, not least because digital channels now account for over 70% of total global ad investments,[1] a historical shift which has accelerated since the covid-19 pandemic. Social media spend forms a significant portion of digital advertising investment, because advertisers want to ensure that they are targeting their products at the places where their customers are most likely to be found, with the broadest possible reach. They also wish to avail themselves of the targeting opportunities which the platforms provide, in order to maximise the efficiency of their campaigns and their return.
2.3. However, as with any media type, with greater use comes higher attendant risks of harm. Given that the platforms do just that – provide a platform for user-generated content – there will be both good and bad actors in the environment. The model of the platforms rewards attention and promotes content which is being more interacted with. This is not always pernicious, and the Committee will be familiar from their own lives with many examples of viral content which might entertain or educate, or promote positive trends such as charitable giving. However, illegal and/or harmful content can also be surfaced by algorithms, necessitating action by government, regulators and platforms themselves.
2.4. Advertisers who utilise these platforms to reach potential customers have a clear interest in supporting the creation of safe online environments, both for themselves as brands and for users, not least because it is their spend on campaigns delivered by the platforms which enable those companies to operate on their current model. To illustrate the scale of this spend, the top 1,000 UK advertisers spent £13.3 billion in 2024, of which £3.9 billion (or 29%) was spending on digital (not including social media).[2] ISBA members represented £5.3 billion of the total spend. Global ad spend was projected to surpass $1 trillion in 2024; and Google, Meta and Amazon are expected to take 70% of the incremental spend between 2024 and 2026, demonstrating the financial disparities (which, when it comes to advertisers, should be looked at from an individual, not collective, perspective).
2.5. ISBA has sought for many years to be proactive in working directly with platforms, with and on behalf of our members, to improve accountability and transparency in the field of online safety. This work became particularly urgent in the wake of the Times’ ‘Brands funding terror’ exposé in 2017,[3] which revealed that brand ads were appearing next to unacceptable content. In the aftermath of these revelations, ISBA worked directly with YouTube to demonetise such content. We have also acted around individual instances of hateful conduct; for example, we convened roundtables for our members with the platforms after the horrendous abuse suffered the England men’s football team’s Black players after the Euro 2020 final. We have also worked with government, for example as a member of the Home Office’s previous taskforce on child sexual exploitation and terrorist propaganda online. Throughout, we have publicly called for proportionate regulation of digital platforms.
2.6. This regulation eventually took the form in the UK of the Online Safety Act (OSA). We regretted the time that this took to reach the statute book (indeed, when it was first mooted, TikTok did not even exist). The Committee will be aware that the OSA is coming into force over the course of 2025, and that Ofcom is working on its implementation. The Act is focused on what are termed the ‘walled gardens’ – closed platforms or ecosystems like Facebook or X where the provider company controls the hardware, access to content, applications, and user data while supposedly restricting access to non-approved items. The OSA does not apply to the broader ‘open web’.
2.7. ISBA and our members welcomed the Act because of advertisers’ wish to operate in an environment which is safe for users and for brands. It has been our consistent position that it is not for advertisers to dictate types of content which can appear on platforms; this is for them, in the same way as newspaper content is a matter for editorial boards. Further, it is for government to set out what is illegal or harmful, and platforms must adhere to that direction and set out as part of their terms and conditions how they will tackle the presence of any such content. On the basis of reporting against those conditions, advertisers can make an informed decision as to whether and how they will place their adspend and run campaigns. To that end, there should be independently audited reporting of the effectiveness of content policy enforcement.
2.8. It has also been our consistent position that any organisation of collective action by advertisers, such as boycotts, would be illegal under competition law. Where we have sought to focus our efforts is on voluntary action to promote brand safe environments, transparency, and accountability.
2.9. The Committee will be familiar with the Global Alliance for Responsible Media (GARM), a voluntary, cross-industry initiative established under the aegis of the World Federation of Advertisers to help address the challenge of defining illegal/harmful content and its potential monetisation via advertising. GARM was set up in the wake of the Christchurch, New Zealand mosque shootings in March 2019, which the killer livestreamed on Facebook. The major tech platforms were members of the initiative.
2.10. GARM’s major output was to develop a Brand Suitability Framework, a taxonomy which defined and described harmful online content in consistent and understandable language. Prior to the Framework’s development, the definitions of harmful content varied by platform. GARM also developed a Brand Safety Floor, which described content which was inappropriate for any advertising support. This included terrorist content, child sexual exploitation, self-harm, animal cruelty, fraud, and piracy. Platforms’ adherence to the Framework and Floor, as well as their compliance with their own terms and conditions on the prevalence of unacceptable content, were assessed in regular, publicly available transparency reports. These tools provided information to help advertisers avoid inadvertently supporting harmful and illegal content.
2.11. In 2024, the Judiciary Committee of the United States House of Representatives undertook hearings and published a report alleging that GARM drove anti-competitive behaviour by disfavouring conservative news outlets in the US. Subsequent legal action in Texan courts led to a statement by the WFA shuttering GARM in August 2024, citing “allegations that unfortunately misconstrue its purpose and activities” and which “significantly drained its resources and finances”.[4]
2.12. These events, alongside the appalling murders in Southport and the subsequent violent disorder, have clearly sharpened the focus on how effectively tech platforms are adhering to their own terms and conditions on unacceptable content and also on what next steps could be taken to bring further transparency and accountability. In the first instance, we believe that the effectiveness of the OSA should be assessed after its provisions are brought fully into force. Should the Government decide to consult on or bring forward fresh online safety legislation, ISBA stands ready to provide expertise and the advertisers’ perspective on any proposed reforms.
What role do generative artificial intelligence (AI) and large language models (LLMs) play in the creation and speed of misinformation, disinformation and harmful content?
2.13. As in every other sector of the economy, generative AI is bringing disruption, challenges and opportunities to advertising and marketing. Clearly, in an online world where programmes which enable the user to create almost any image they wish via a series of prompts, and then release that content on to social media platforms with wide reach, the question of potential harm becomes acute – and brings the importance of adherence to terms and conditions on unacceptable content into even sharper focus.
2.14. Regulation of Gen AI is in its infancy, but we believe that guiding principles being set down by our industry indicate advertisers’ thinking. In November 2023, ISBA and the Institute of Practitioners in Advertising (the IPA, representing ad agencies) published 12 guiding principles on the use of Gen AI in advertising, including (but not limited to) content.[5] These are broad-brush, and are designed to help the industry to embrace AI in an ethical way that protects consumers and those working in the creative sector. The principles include preventing AI’s use in a way likely to undermine public trust in advertising (e.g. deepfakes or scam ads), avoiding prejudiced or discriminatory outcomes, the importance of human oversight, and transparency with the public when AI has been used and is unlikely to be obvious to consumers. These principles of course deal with legal content and are self-regulatory, but they demonstrate industry’s commitment to and expectations of the platforms when it comes to their terms regarding the deployment of AI in user-generated content.
How effective is the UK’s regulatory and legislative framework on tackling these issues?
2.15. As noted above, we welcome the progress which Ofcom is making on publishing guidance under the Online Safety Act. We do believe that this should have an impact on the prevalence of harmful content and have noted the commitment of the platforms to complying with local law.
2.16. The Committee will recall the debates over the OSA and the changes which were made under the previous Government, ostensibly in order that free speech was not curtailed; adult safety duties on ‘legal but harmful’ content were removed from the Bill as it proceeded through Parliament. In the place of those duties, platforms were required to provide optional user empowerment tools to give adult users greater control over the content that they see, including a list of content categories set out in legislation. Hate speech is listed as among the priority offences. When it comes to children, the Act requires that platforms must remove a series of categories of content that is harmful or age-inappropriate, including those which have been raised in oral evidence sessions during this inquiry such as self-harm and eating disorders. We believe that this underlines the importance of assessing the effect of the Act as it comes into force.
2.17. Misinformation and disinformation are captured by the Act in so far as their content is illegal or harmful to children, or if the companies defined as Category 1 services under the legislation have prohibited types of mis/disinformation in their terms and conditions.[6] Ofcom has established an advisory committee to look further at understanding mis- and disinformation, which will have its first meeting in the coming weeks. As part of this work, attempts will have to be made to distinguish between mis- and disinformation disseminated with unlawful intent – such as that in evidence last summer; potentially ill-informed opinion; or simple misunderstanding which may go viral. ISBA will of course engage with Ofcom and government on this work and any potential new online safety legislation which may result.
2.18. We have pointed out that the OSA applies to the walled gardens, but not the open web. Advertisers wishing to place ads on the broader internet navigate a complicated and opaque set-up of programmatic advertising which contains a plethora of actors in the supply chain. From the moment the advertiser commits their spend, that money will pass through agencies, demand-side platforms and then supply-side platforms before it reaches the publisher. These transactions can take fractions of a second as ad space is automatically auctioned before an ad appears. ISBA has conducted studies with advertisers and publishers – notably, our Programmatic Supply Chain Transparency Studies in 2020[7] and 2023[8] – which noted the ‘unknown delta’ of spend that could not be accurately traced through this system. It also identified the average of number of websites that would be included in a campaign - 9,000 in the latest study. This number may include so-called ‘Made for Advertising’ (MFA) sites which are created for the singular purpose of simultaneously buying and selling ad inventory; typically, they use sensational headlines, clickbait and provocative content to attract visits and generate page views, which in turn generates ad revenue for the site owner. Such sites are often designed to fool digital advertising buyers.[9]
2.19. The programmatic system has evolved rather than been designed, and is still relatively young, programmatic ads only having first appeared in 2007. Its subsequent rate of growth has been exponential. For the Government to legislate or regulate in this broad space, in the face of a programmatic system which is so entangled, would be a substantial challenge; ISBA and industry partners have previously worked to promote standardisation to help facilitate data sharing so that spend can be audited end-to-end. Such increased visibility would help marketers to ensure that their ads are not appearing in places likely to be promoting unacceptable content and/or driving harm.
2.20. One potential route to increased accountability in the programmatic environment, which has been put forward by the UK Stop Ad Funded Crime Group (UKSAFC), is to impose Know Your Customer obligations on participants in the programmatic supply chain. This has the potential to combat both fraud and the spread of misinformation through MFA sites. The proposal would introduce considerable friction in the supply chain and, in ISBA’s view, would need careful evaluation and consultation before progressing.
2.21. The enforcement of the Online Safety Act is currently the primary vehicle for the tackling of illegal content and content that is harmful to children. The UK enjoys a world-leading and respected system of advertising self-regulation, enforced by the Advertising Standards Authority (ASA) – but its remit is legal advertising. The ASA has worked to ensure that its writ runs wider with the platforms via its Intermediary and Platform Principles (IPP) workstream, which has explored putting on a more formal footing, and bringing consistency to, the ways in which those companies co-operate with the regulator to promote advertisers’ awareness of the rules online, and to help the ASA secure compliance in cases where an advertiser refuses to amend or withdraw an ad.
2.22. Meanwhile, the Government’s Online Advertising Taskforce – of which ISBA is a member and within which we chair two sub-groups, on age assurance and influencer marketing – is focused on harmful advertising, particularly to children and young people.[10] The Taskforce is working to improve the evidence around the scale and threat of the in-scope harms; and to identify ways to enhance voluntary industry initiatives or standards to tackle their drivers. It will report on progress in the autumn of 2025. Both the Taskforce and the ASA IPP work are focused on the content of advertising rather than its process, and each initiative has different but related ends in mind.
***
Contact: Rob Newman, Director of Public Affairs, ISBA
6