Written evidence submitted by Professor Kevin McDonald (SMH0020)
Social media, misinformation and harmful algorithms
Introduction
I am professor of Sociology at Middlesex University in London. My broad area of research is social movements and social mobilisations, and my recent work has focused on violent extremism, radicalisation and terrorism. I publish in this field, my most recent book being Radicalization (Cambridge, Polity Press, 2018). This book focused on the narratives and experience constructed by young people who attempted to travel to Syria to join the Islamic State group between 2013-17, the research based on interviews and the analysis of social media posts of these young people in the period leading to their departure.
Over the period since August 2024, I have been analysing the social media posts of participants in the riots that took place during the summer.
My contribution to the Committee’s work focuses on two areas:
Algorithms play a vitally important role in constructing communities where it appears that ‘everyone’ agrees around core ideas and feelings (eg outrage, anger, frustration etc).
It is very important to understand that this is not simply a matter of ‘misinformation’. This is term from the physical sciences, and is not always appropriate to understand social phenomena. The affective structure of communications (ie the feelings it is built around) are central to the development of extremism and violence. Algorithms are not to be understood in terms of ‘cause’ and ‘effect’, but they play a vital role in the construction of rumour and conspiracy theory. These were at the heart of the violence that took place over the summer. This is evident in the focus on children and the ‘save our children’ themes, which were not at all focused on the proposition that there are too many migrants in the United Kingdom, but on the proposition that refugees pose a threat to children.
This is a complex and ambiguous theme but has been at the heart of violent protests against asylum seekers – as in Merseyside in February 2023.
While algorithms play an important role in creating communities that are sealed off from others, they are not the source of rumours and mobilisations. Social media is participatory media, as we can see by the number of violent protesters filming and streaming events to online platforms. Such events are amplified not only by algorithms but also commercial media, in particular television and radio.
An increasing proportion of the UK population receives news primarily, or exclusively, from social media. Much of this content is produced by activists of different kinds, and this is particularly the case around asylum seekers, where a generation of ‘citizen journalists’ are involved in campaigns against the housing of migrants, and claims about the risk they pose to children.
Algorithms are very important in curating the media feeds that an increasing proportion of the UK population receive. Increasingly these aim at maximising ‘engagement’, and this in turn involves recommendations of content that is most entertaining, disturbing, or horrific. This has clearly had the effect of communicating claims about the supposed menace that asylum seekers pose to children, and in mobilising the kind of violence seen during the summer of 2024.
This framework addresses crisis events and aims at rapid intervention. This is important and welcome. However, there is a form of ‘background hum’ associated with conspiracy theory, online videos etc, that involves the ‘dehumanisation of the other’. This can take the form of humour, stories of horror, etc. The regulatory framework is less adept at identifying and addressing this, in particular when these communications are produced by ‘citizen journalists’, meme groups, gamer platforms, etc.
To the extent that addressing increasing forms of polarisation and dehumanisation is a ‘whole of society’ challenge, there would be significant benefit in addressing more clearly how ‘hate’ works as a social process, and exploring ways that government action can play a role in empowering local communities to counter the kind of apocalyptic communications framed in terms of ‘great replacement’, ‘they are raping our children’, etc.
The Italian example?
In Italy there is an important national policy since 1996, where the proceeds of state confiscation of assets of mafia and other organised criminal actors do not go to state coffers, but the funds obtained are used to support local economic and social initiatives. This has had a transformative effect in those regions most blighted by organised criminality.
In the UK, where the regulatory framework imposes penalties on media companies for permitting communications that promote hate and violence, there is significant scope for the funds obtained to be directed to supporting community-building communication initiatives, modelled on the principles shaping the Italian legislation (Law 109, 7 March 1996). Such initiatives should be considered as measure to repair the damage caused by hateful communications.
This suggests a regulatory model that extends beyond controlling the negative consequences of social media and extending supporting forms of community recovery and empowerment in those areas most damaged by hateful and extremist communications. Such an initiative would be an important measure towards supporting a ‘whole of society’ response to harmful and false content online.
17 December 2024