Written Evidence from Mr Jacob Rowbotton (DFF0016)

1. The Committee is rightly concerned about reports of harassment, abuse and intimidation directed at politicians and others in the public eye. Before considering any action to tackle this problem, it is worth putting the current trends into a broader context. Twenty years ago, commentators regularly complained about politics becoming too stage-managed, in which political actors exercised too much control over the terms on which they engaged with the public. The current culture has moved away from that state of affairs and the current concerns reflect a loss of control that political actors have over publicity. To some, the issues surrounding abuse may represent the darker side-effect of a broader democratisation of communication.

2. When hearing reports of abuse, intimidation and harassment of political actors, it is tempting to seek new laws, call for stronger penalties or demand stricter enforcement of existing laws. Before any such action is taken, it is important to keep in mind a number of challenges. Most obviously, it can be difficult to separate genuine exercises of free speech (albeit in a heated context) from acts of abuse. A further challenge is that abuse and intimidation can be used to describe very different patterns of behaviour, which may call for different responses. Before proposing any concrete changes, it is important to identify which patterns of behaviour are most problematic. Any future regulator of social media could play a role in acquiring further information, as well as devising appropriate responses from the social media companies to address the problems. 

Speech rights

3. When considering the balance to be struck between freedom of expression and other competing rights, there are special considerations where the communication relates to a politician. It is a well-established principle under the ECHR that political figures and those in public office can be expected to tolerate greater criticism than a private individual.[1] This can have several legal implications. For example, a disclosure about a politician is less likely to be protected in privacy law.[2] When interpreting criminal laws regulating abusive messages, the courts can accommodate the free speech rights by raising the threshold of harm for the offence, or when applying a defence of reasonable conduct (where applicable).

4. While a political figure is expected to tolerate more robust criticism, that does not mean there are no limits.[3] The Article 10 jurisprudence offers relatively limited protection for gratuitous insults or hate speech. In some cases, however, there will be difficulties in applying these broad standards - for example where a term of abuse (suggesting lower Art 10 protection) is used in the course of political discussion (suggesting stronger Art 10 protection). While much will depend on the context of the statement, many past cases dealing with abusive messages in a political context have been criticised for failing to strike an appropriate balance.[4] It can be particularly difficult to work such a fact-sensitive balance into the guidelines and policies for law enforcement, which aim to ensure some consistency in application of the law.

Persistent abuse and intimidation

5. As the Committee’s call for evidence makes clear, persistent hostility and abuse can deter people from speaking in public or on publicly-accessible forms. The issue can go beyond deterring the exercise of speech rights, as persistent and extreme forms of abuse could in practice limit a person’s right to stand for election (which is a separate right).[5] Such an impact is clearly relevant in striking a balance, and can provide a basis for the restriction of (the abuser’s) expression rights under Article 10(2).[6]

6. It can nonetheless be difficult to identify when a message or communication has a detrimental impact on the rights of the politician. For example, a communication (or prospect of a communication) that deters a person from seeking public office does not necessarily constitute intimidation. A person may decide not to stand for office because they do not want to be the subject of media scrutiny and criticism (which fall within the boundaries of legitimate debate). Sometimes members of the public may seek to dissuade a political candidate from seeking public office on the grounds that the candidate has a track record of incompetence. While the speaker in that example may intend to deter a person from standing for office, that would normally be considered legitimate criticism.[7] Something more is required for an action to constitute intimidation – namely that it does not seek to persuade or make a person aware of the reasonable consequences of their actions. Instead intimidation makes a person fearful of certain consequences (that could not be considered legitimate) if they pursue a particular course of action.

Different patterns of behaviour

7. The abuse of politicians through social media is often discussed as a general problem. However, terms such as abuse and intimidation can be used to describe a range of different types of behaviour, which may call for a different response.[8] Consider the following examples:

(i)               A person sends a message to a candidate that credibly threatens serious injury to members of her family.

(ii)             A person sends a candidate hundreds of messages over a period of weeks that are insulting or abusive (but which do not contain a threat of injury).

(iii)          A person posts a comment on a news story on social media that is available to the world at large, flippantly saying he wishes someone would shoot a particular politician.

The first example is the most straightforward and likely to be caught by the existing laws.[9] In example (ii), the individual messages taken in isolation may fall within the boundary of permissible criticism. However, it is the aggregate effect of the multiple messages that can cause serious distress and disturbance to the politician, and this example may fall within the law of harassment. The third example is more complex as the message is not directed or sent to the politician (even if the politician sees it). Instead, it is a general comment to the world at large and is less likely to cross the threshold set in the criminal law.[10] It is also most likely to be regarded as political speech (in so far as it is commenting on a matter of general interest). The presence and widespread availability of such comments may nonetheless be distressing to the politician.

8. The above examples indicate several different types of abusive behaviour. In those examples, the messages are assumed to come from one source. Any potential liability can thereby be assessed in relation to that person’s actions. There are, however, problems of scale in attempting to impose legal sanctions on such behaviour. There are not the resources to investigate every abusive message that could potentially fall foul of the existing laws. The police and prosecutors therefore have to develop internal rules and policies to prioritise resources, while also ensuring that the laws are applied consistently and that expression rights are respected. This means that, even with appropriate laws in place, there are rules of thumb guiding enforcement that may not always be sensitive to the context.[11] In so as far as is possible, it would be useful to acquire information about the policies that guide law enforcement on intervening in the context of political disputes. By this, I do not mean to criticise the existing practices, but merely illustrate that the law can be a blunt tool for addressing this type of problem.

De-centralised attacks

9. A further challenge arises where the impact on a politician comes not from the actions of a single individual but from many people. To illustrate the point, modify the example (ii) above so that the insulting or abusive messages come from hundreds of different people (none of whom send more than one message).[12] Even if no individual thereby publishes content that crosses the threshold in criminal law, the effect on the politician may be just as great. In its recent review, the Law Commission concluded that it is not clear how well the existing law deals with such collective behaviour.[13]

10. Sometimes such widespread activity may be truly decentralised and uncoordinated. In such a case, individuals make an independent choice to post an insulting comment or message (maybe inspired by seeing others post similar messages publicly on the social media). However, in other cases, there may be greater coordination. A group of people may decide (either offline or in another forum) to act in concert with one another to post comments or send messages to an individual.[14] Another example is where a person or body employs a ‘troll army’ to insult or undermine an individual, generate a hostile environment for certain viewpoints, or generate confusion about an issue. While the messages appear to be from lots of different individuals, they are acting as agents for the same person or body. Distinguishing these latter types of activity is important, as they cannot simply be characterised as part of the rough and tumble of politics, but undermine the workings of the political system in so far as the person behind the messages is unknown. However, the level of coordination may not be apparent to the person targeted and be harder to detect.

11. Other smaller steps could be taken that do not impose liability on a publisher, but which seek to make the online environment less hospitable for abusive content. For example, the CSPL (along with the Electoral Commission) recommends extending the imprint requirements on printed electoral material (showing who is responsible for a message) to be extended to online material and the social media. If this can be successfully implemented, it might help address problems with certain anonymous electoral messages (though it will not catch all anonymous messages). Another important step is the protection of privacy (whether through existing laws such as the Data Protection Act 2018 or through future social media regulation). Online intimidation can include the publication of private details (such as disclosing a person’s home address or publishing photographs of a person’s home and family members). Provisions requiring the social media companies to takedown posts disclosing private information can help to remove one of the tools frequently used to intimidate people online.

The role of social media companies

12. The social media companies can play a central role in addressing the abuse of politicians online. At the time of writing, proposals are being considered for social media regulation. In particular, the Government’s proposals are likely to require social media companies to have systems in place for dealing with certain types of abusive content. There is much sense in this strategy. Requiring social media companies to take action will be more efficient than targeting the numerous publishers, and may provide a more proportionate response to some of the problems. While a person should not attract a legal penalty for posting certain types of political message, that does not mean the person has a right to have that message available to the public indefinitely or turn up in search results against a person’s name. Under this approach, the person may have the right to speak, but the social media company has a requirement to mitigate the harm.

13. It is likely that the abuse of political figures will be covered by any general social media regulations in relation to content that is abusive, defamatory, or incites hatred (although there may be a provision requiring a balance with political speech). In addition, social media companies could be required to follow codes of practice in relation to electoral messages. This could include requirements on the transparency of electoral advertising, and the presence of systems to detect foreign interference or other coordinated activities.

14. As concrete proposals have yet to be published, it is difficult to say how effective future regulation will be (and whether there are any drawbacks). However, aside from codes of conduct and content policies, any new regulator could work with social media companies to assess the prevalence of abusive messages directed towards politicians (as at present much evidence is anecdotal). The regulator could seek to identify patterns of behaviour, such as whether reported abusive messages tend to be sent at certain times of day, or are associated with certain events.[15] This might help in understanding what leads a person into such behavior, and whether social media companies can develop methods that make it less likely for people to engage in actions that intimidate others.[16] More broadly, a new regulator could require some assessment of (and improvements in) the systems for reporting content or blocking people.[17] The social media companies could also be asked to consider whether the algorithms tend to reward extreme and abusive content, and whether steps can be taken to address such a problem. That is obviously more complex than passing a ban on certain types of speech, but is more likely to address the root causes that incentivise abusive behavior.

25 March 2019