Elections Bill

Written evidence submitted by Dr Katharine Dommett, University of Sheffield (EB01)

Research expertise in digital campaigning and transparency

This submission of evidence relates to Part 6 of the Elections, focusing on the proposals to introduce digital imprints.

Summary of key points:

- The Bill stipulates that an imprint does not need to be displayed on material ‘[i]f it is not reasonably practicable’. This creates potential challenges in implementation as the bounds of what is ‘reasonably practicable’ are not defined. As such there is potential for campaigners to interpret this provision in a range of ways, resulting in large amounts of material not displaying a imprint on the material itself (with this content hosted elsewhere). This would create challenges for scrutiny and research and would affect public access to this information – frustrating legislative goals. As such there is a need for clearer guidance of what ‘reasonably practicable’ means in practice (for more see point 4 below).

- The Bill makes a distinction between paid and unpaid material which has three consequences.

1. In specifying paid content, the Bill focuses on certain types of payment – namely, payment for dissemination. This would lead to some types of activity not being subject to imprints (i.e. influencers – who are paid for content creation, not distribution, which is done organically). To advance the goals of the imprints - to improve transparency around the origin and financial source of campaign material – payment should not be measured only in terms of distribution, but also content creation. (for more see point 7 below)

2. The distinction between paid and unpaid potentially creates confusion around requirements for campaign material when it changes content type (i.e. when originally paid material is subsequently disseminated organically, or vice versa). This suggests that the distinction may be difficult to enforce in practice, and at least suggests the need for additional guidance on what happens when materials change format. (for more see point 7 below)

3. The distinction is accompanied by different regulatory requirements, specifically excluding unregistered third party campaigners from including an imprint on unpaid campaign material. Given evidence of large scale organic campaigns, which can often be coordinated, this could result in a swathe of election material lacking transparency information. For this reason I recommend that all forms of content should be subject to the same requirement to carry an imprint (for more see point 8 below)

Detailed points:

1. In relation to point 35, 3(b) it is notable that the provision does not extend to SMS text messages. This is a potentially significant gap given that many text messages are accessed on electronic devices and through applications (such as WhatsApp). Whilst not commonly used in the UK, the use of SMS messages in campaigns has increased dramatically in the US and Brazil, but such content would not be subject to oversight.

2. In relation to point 35(4) it is indicated that ‘"The promoter", in relation to electronic material, means the person causing the material to be published’ – the phrase ‘causing’ here could be possibly misleading. I.e. if an advertising intermediary is used – such as SmallAxe Communications– they would be ‘causing’ the advert to appear in a material sense, but they are not the actor responsible for conceiving that paid content. As an alternative I suggest that, in line with 38(4), the language of ‘paying for’ the material to be published would be more appropriate here.

The need for a clear differentiation here is required as unlike in PPERA (126 (2)) where there is a stipulation to include the printer as well as the promoter – providing a point of differentiation which implies what is required in stating the promoter – this is not present. This does raise the question as to whether there is a case for compelling the disclosure of any ‘intermediary’ actor – such as an advertising agency or consultancy firm – who is placing the advert on behalf of the promoter.

3. On 35(6) the need for amendment power will be important to allow future updates, particularly developments in mixed-media such as hybrid forms of television/mobile advertising or smart speaker adverts are adopted.

4. On point 37(2) the stipulation about including an imprint unless ‘it is not reasonably practicable’ creates a potential opportunity for campaigners to avoid placing an imprint on digital material. It’s not clear how this will be interpreted and what therefore constitutes a reasonable practicability. It is also likely that allowing campaigners to place imprints elsewhere will create barriers for scrutiny and research – as it will be necessary to track down where information is listed. It also lessons the transparency advantage for the public as they are unlikely to click for further information.

5. On 37 (5) – this provision outlines the need for text and image imprints, or audio imprints to match the type of electronic material in question (i.e. audio campaign material needs an audio imprint). The stipulations leaves ambiguity about what form of imprint is required when both forms of media are in evidence. This could occur, for example, where an advert contains a series of changing imaging that relate to the campaign theme – i.e. a picture of a flag, protests, party emblems – and also contains an audio overlap that promotes an electoral message. In such situations it is not clear which type of imprint should be used, or whether both should be in evidence. Given that some visual ads with audio do not auto-play the sound, there is a case for text being primary – however it is possible for audio first systems to exist – i.e. podcasts that have an accompanying image deck. This indicates a case for both types of imprint to be required on mixed-media campaign material.

6. On paid for material, It is worth noting that the criteria outlined in 38(3) would likely not require mobilisation/ registration campaigns to contain an imprint because the wording emphasises material with the purposes of ‘influencing the public, or any section of the public, to give support to or withhold support from… a registered party) etc. This could suggest an additional statement such as ‘the electoral process’. This could potentially help the identification of voter registration and suppression campaigns which can, indirectly, assist party/ referendum etc campaigns.

7. On the distinction between 38 (paid)/39 (unpaid) material. In general there are reasons to support this distinction and, indeed, it’s something that at the level of research is useful. However, thinking about enforcement further and the gaps this distinction creates I’m no longer convinced it’s a useful foundation for differentiation within the Bill.

This is because the focus on distribution means that any money expended on content is not factored into this regulation. The implications of this are evident when thinking about influence campaigns – which are becoming increasingly significant.

An influencer is paid for their organic content and reach, but they themselves rarely pay for distribution of their message. Under these rules, they would not be required to carry an imprint as they are most commonly unregistered third party campaigners and are not placing paid content. This creates a significant gap in regulation and a lack of transparency around the role of these actors.

In addition, it creates potential ambiguity about content that changes classification. I.e. in my evidence to the CO I highlighted how materially that was originally unpaid, could subsequently be ‘boosted’ via payment, and similarly paid material can be subsequently shared organically. The lack of a consistent framework here creates ambiguity about what should happen in these situations where content originates from unregulated third party campaigners.

There is precedent in Scotland for paid and unpaid material to be subject to the same imprint requirements, and I am of the view that this is now preferable to capture such examples.

8. On 39, this provision indicates that unpaid content from unregistered third party campaigners will not be subject to regulation (i.e. of unpaid content). This provision was presumably designed to ensure that public opinion was not regulated, however it may not be the most appropriate route for regulation.

Within Scotland paid and unpaid content from unregistered third party campaigners are subject to imprint regulation, but there is an exemption for expressions of public opinion. I believe this provides an important way forward and would allow for a more comprehensive and therefore potentially less confusing regime.

If unpaid content from unregistered third party campaigners does go unregulated then it is important to recognise that campaigns designed to occur ‘organically’ – which are becoming an important tool for campaign messaging – would not include transparency mechanisms. This is significant as this tactic is often used, meaning much content would not contain this provision. It is also notable that unregistered groups can co-ordinate such campaign activity and therefore have the capacity to reach a wide audience – especially when their content is favoured by platform algorithms.

9. On the purposes of unpaid electronic communication (40), as above in relation to my sixth point, the focus on the success of specific actors precludes campaign activity focused on boosting or supressing participation. This again may suggest a case for including a wider focus on electoral participation messaging, not simply on the actors involved in securing a particular electoral outcome.

10. On point 42(3) which provides an exclusion for material that is published on a website or mobile application whose primary purpose is journalism. This creates an area of ambiguity as there is a lack of specification as to what constitutes a journalistic website/ outlet. Would, for example, partisan ‘news’ sites set up by campaign groups be exempt from carrying imprints? And would social media sites – where increasingly numbers of citizens consume news – see ‘news items’ exempt from carrying an imprint? Given that previous research has shown campaigners to utilise news features to circumvent campaigning restrictions, there is a need to more clearly specify what constitutes a journalistic outlet.

11. On enforcement 44(2) contains provision for removal of content by the original publisher, however this provision does not allow for removal of the same content which has been re-published on other platforms. Extending this provision to allow the removal of all versions of the image would ensure that this was a more effective disincentive. Otherwise, given the notice period given to promoters, there would be ample time for them to replicate the same message on other platforms.

12. Within the provisions on enforcement, as outlined in the Cabinet Office’s technical consultation response there is an apparent discrepancy between the size of fines that can be levied, as material supporting candidates can receive an unlimited fine, whereas material supporting parties only a £20,000 fine. This suggests a case for aligning fine levels.

13. On point 47 (4), the requirement for a minimum notice period of 14 days in which a promoter can make representations against a take-down notice may allow problematic content placed online immediately before the election to remain in place. There would accordingly be little disincentive to placing content not conforming to these rules online in the 14 days before an election – the very period in which most citizens are attentive to political messaging. This suggests that a reduced time limit and enforcement process may be required in the immediate days before an election.

14. The Bill does not give the Electoral Commission further power to compel information to allow their enforcement activity. At present they do not have powers to compel information from third parties such as digital platforms or unregistered campaigners unless they suspect an offence has been committed and open an investigation. It would be advantageous for the Electoral Commission to have additional powers so they did not need to open a formal investigation in order to contact an individual to amend an imprint (something informally done with registered campaigners/parties).

14 September 2021