AI chatbots are more likely to source answers to questions about the news from Left-wing sources such as the Guardian, according to a think-tank.
Analysis from the Institute for Public Policy Research (IPPR), published today, reveals the popular tools used by millions are drawing on a 'narrow and inconsistent' range of sources.
They found that the BBC was the most cited source in Google AI Overview answers, appearing in 52.5% of answers.
The Guardian came top for ChatGPT and Google Gemini, appearing in 58% and 53% of answers respectively.
Meanwhile, the Daily Mail, alongside other right-leaning publications, appeared in only a small fraction of the answers – despite having a substantial combined market share of the UK audience.
The findings have raised concerns that this editorialisation by AI companies is creating a new generation of 'winners and losers'.
IPPR experts warn that the source selection risks narrowing the range of perspectives users are exposed to, potentially amplifying particular viewpoints or agendas.
Roa Powell, senior research fellow at IPPR, said: 'AI tools are rapidly becoming the front door to news, but right now that door is being controlled by a handful of tech companies with little transparency or accountability.'
AI tools pay for and prioritise some outlets, exploit other content for free, or even exclude sources that block access.
Some publishers, such as the Guardian, have licensing agreements with firms such as ChatGPT, while others – including the BBC – have sought to block AI companies from accessing their content.
But AI tools do not inform users about why sourcing decisions are made, meaning users often have no idea about the editing that is going on behind the scenes.
Researchers also argue the rise of AI could have serious consequences for the financial sustainability of quality journalism.
For instance, when a Google AI Overview is present, Google users are almost half as likely to click on news links to news publishers - threatening advertising and subscription revenues.
Owen Meredith, chief executive of News Media Association, said: ‘As the Report demonstrates, weakening UK copyright law would deprive publishers of reward and payment for the trusted journalism that enables AI to be accurate and up to date.
‘The CMA must intervene swiftly to stop Google using its dominant position to force publishers to fuel its AI chatbots for free.
‘Fair payment from the market leader is critical to a functioning licensing market and to preventing big tech incumbents from monopolising AI.’
Carsten Jung, associate director for economic policy and AI at IPPR, said: 'So far, much of AI policy has sought to accelerate AI development.
'But we are coming to a stage where we need to more deliberately steer AI policy towards socially beneficial outcomes.
'In the news space, we have the tools to ensure that AI does not damage the public sphere, and in fact improve the quality and diversity of information people access.
'But this won't happen by itself – the government needs to shape it. We should learn the lessons from the past and shape emerging technologies before it is too late.'
