AI is putting pressure on democracy
Theme: Tech & Democracy
AI makes it easier than ever to create credible but false content – and this is having an impact on politics. Anders Hedman, a researcher in media technology, warns of increasing polarisation, declining trust and a democracy under pressure from rapid technological development.
Anders Hedman paints a critical picture of how AI tools on social media platforms have developed – and argues that this has created problems for both politics and elections:
Since the 2010s, legally protected social media platforms, through technology designed to influence our behaviour, have contributed to increased polarisation and influenced election debates. With AI, this has become even more powerful – today it is easy to create credible but manipulated content on a large scale.
Algorithms also drive polarisation as a consequence of being designed to maximise the time we spend on media platforms with content that is emotionally charged and sensational. The result is that loud political opinions spread faster than measured and fact-based ones, posing a threat to democracy, argues Anders Hedman.
– The companies behind the platforms are not making any significant effort to stop this, but are focusing on strengthening their positions in the media world. At the same time, more falsehoods are being spread as if they were facts, which is hardly good for democracy.
Responsibility for the content
Anders Hedman researches and teaches on AI, human–computer interaction and the impact of technology on society and human thinking. Together with the Journalist Fund, he has developed and delivered a KTH course for journalists on AI.
– AI has quickly become part of how news is produced, disseminated and received. But it is not just about understanding new tools and working methods. Journalists also need to understand the technical and economic power structures within which AI systems operate, as well as the possibilities and limitations of the technology itself, he says.
Another issue highlighted in the course is the difference between social and traditional media. Social media platforms are protected by freedom of expression laws, but without the same responsibility for content as traditional media with responsible editors.
– At the same time, AI-supported influence design on social media offers significant opportunities for political and other actors to mislead and manipulate opinions.
Anders Hedman believes that a key aim of the course is to give journalists a better understanding of AI, so that they can report on the subject credibly and not get swept up in all the hype.
AI and human intelligence
He also calls for a broader and freer discussion about where AI is actually heading, preferably from a more philosophical perspective. Today, he feels that the debate is often steered from within the AI world itself, where, for example, AI is compared with human intelligence.
According to Anders Hedman, there is an “AI elite” that views human thinking as purely computational processes and intelligence as achieving goals in efficient ways. He believes this is both simplistic and misleading. He warns that this technocratic view can be dangerous.
– If we start to see AI as the yardstick for intelligence and creativity, we risk losing our own ability to think innovatively and see things from different perspectives.
At the same time, he sees great potential for AI, not least in journalism. But this is based on having proper knowledge within one’s field.
– Then AI can become a tool for sharpening analysis, finding new angles and making smarter connections – something that fundamentally applies to us all.
Text: Christer Gummeson ( gummeson@kth.se )