As EU AI Act enters into force, study warns chatbots blur voters’ own political views
by https://euobserver.com/author/gaia-neiman/ · EUobserverThe LSE study found political interactions sometimes become problematic since chatbot replies decide and filter result rankings according to its Western-heavy training, thus underrepresenting the views of minorities. (Source: Unsplash)
Unlock article and share
By Gaia Neiman,
Brussels
,
The EU’s artificial intelligence act came into force on Sunday (2 August), while new research reveals how AI is eroding citizen engagement in democracy.
In a paper published on Monday (3 August) in Perspectives on Politics, researchers from the London School of Economics (LSE) found that the effects of AI on our understandings of politics go far beyond misinformation and deepfakes.
The study found that when users generate political content through an AI chatbot, it becomes harder to identify whether the political opinion is their own.
Using an AI chatbot to understand a policy or electoral candidates means its active involvement in how citizens interpret the issue in the first place.
The EU has kept a wary stance on AI due to algorithmic bias in its back-end, promoting political actors favourable to those running the chatbots.
The LSE study found political interactions sometimes become problematic since chatbot replies decide and filter result rankings according to its Western-heavy training, thus underrepresenting the views of minorities.
The new law entering into force on Monday regulates AI systems’ transparency by treating their political influence as identifiable, although results from citizen interactions with AI chatbots such as ChatGPT are now shown to weave new opinions in with the user’s existing politics.
According to Eurostat, one of every three EU residents aged 16-74 used generative AI tools in a three-month period before the survey was conducted in 2025.
In a race to get ahead in tech sovereignty, the EU has invested significant resources in its race to become an ‘AI continent’, announcing the construction of up to seven Gigafactories for AI infrastructure worth €30 billion on Thursday (30 July).
“If citizens can no longer tell where their own reasoning ends and the system's influence begins, that changes what it means for a democratic institution to represent the will of the people,” study co-author Elena Pro told EUobserver.
The AI act works to guarantee the safety of AI use across the bloc, entering into force alongside transparency regulations that will require AI systems to tell users when they are interacting with AI or when content has been generated or altered by AI.
Unlock article and share
Latest from AI
EU to co-fund seven AI Gigafactories in race for tech autonomy
EU-banned facial-recognition cameras made in Spain now watch millions of Indians
Latest from Digital
EU risks missing digital jobs target by five million, despite rapid tech growth
Chip shortages forced China to find way to make cheaper AI models than US
Latest from EU politics
Ceuta border surge shakes Morocco’s image as the EU’s ‘trusted partner’
A purge against Orbán-era billionaires could begin this autumn (Day 80 of the Tisza government in Hungary)
The LSE study found political interactions sometimes become problematic since chatbot replies decide and filter result rankings according to its Western-heavy training, thus underrepresenting the views of minorities. (Source: Unsplash)
Topics
Author Bio
Gaia Neiman is a migration and politics reporter at EUobserver. She has previously written for Reuters, The Guardian, The Telegraph, among others.
+ Follow author by email