Building a public interest approach to AI in the news

Developments in artificial intelligence (AI) are changing how journalists make the news and how audiences consume it. From translation to document analysis to writing assistance, news organisations are increasingly using generative AI to research and produce reports. Audiences are starting to use chatbots, AI assistants, and a new genre of AI-enabled search in so-called ‘answer engines’ to get the news and information they need. In fact, they often have little choice about whether they use these AI features, which are increasingly baked into consumer software.

With these changes come new forms of misinformation and disinformation, errors, inaccuracies, and public concern about bias, representation, and the ills of automation. They also bring risks to the sustainability of news providers as AI products draw audiences away from original content, without investing back into journalism.

It is clear AI is reshaping the wider information ecosystem we all inhabit. Technology platforms and AI companies are growing their influence over the infrastructures that underpin how we communicate and what people get to see and hear. Motivated largely by profit-driven engagement metrics, these companies have no obligation to truth or accuracy.

Although it’s easier for people to access news than ever before – it can be harder to wade through the mass of information out there and work out what and who to trust.

What role can public service media play?

As CEO of BBC News Deborah Turness has warned, the price we pay for AI “must not be a world where people searching for answers are served distorted, defective content that presents itself as fact”.

UK public service media aim to help people navigate the digital world to find reliable and impartial news, and better understand the society they live in. Given the challenges AI poses, this is easier said than done.

President of the European Broadcasting Union, Delphine Ernotte, hit the nail on the head when she said: “To make AI work for everyone, we need collaboration – by media outlets, the public, policymakers, and tech companies – and plenty of positive action."

Our team at the Responsible Innovation Centre for Public Media Futures (RIC) is exploring what public service media needs to harness AI in the public interest and foster a healthy information ecosystem. We’re identifying where research insights and interventions can make a difference.

We have found extensive research and development work under way, particularly in the big-ticket areas of tackling misinformation and disinformation, fact-checking, (deep)fake detection, AI skills development and applications of large language (image, and audio) models.

But we also identified important questions that are neglected, approached from an overly narrow perspective, or have in the current climate taken on increased importance.

A critical research agenda for AI and journalism

Our report outlines five key areas that are urgent and necessary to address the emerging challenges. They are about understanding how AI intersects with public service media values, goals, and obligations. They also relate to understanding what people want and need from public service media in an increasingly AI-mediated world.

• AI evaluation that measures what matters in public service media journalism. Newsrooms are integrating AI, but there aren’t yet good ways assess the value this brings to journalism and public service media, or the impact it has on journalists. This makes it difficult to evaluate the effectiveness of AI initiatives and guide good strategy and decision-making about future investments. A fundamental question here is: How do we evaluate the integration of AI in the newsroom against journalistic, public service media and democratic goals and values? Progress here will require the development of new frameworks, methods, and tools.

• Strategic independence amid AI infrastructures and integration. To deliver their mission in a digital world, public service media have to engage extensively with third party technology companies and use their tools and platforms. How can we better understand the implications of relationships and partnerships with AI and (big) technology companies? What does this mean for editorial autonomy and strategic independence of public service media? This will require empirical mapping of existing dependencies and conceptual/theoretical work to interpret the impact they have organisationally and structurally.

• Creating a culture of responsible AI in newsrooms. Public service media are facing challenges as they work to ensure ethical, editorially appropriate, and value-aligned integration of AI into the news production. What can be done to ensure responsible AI in news production beyond just training staff? This will require devising approaches to socialise and embed responsible practices in relation to AI and ensuring editorial and technical processes co-evolve.

• Community engagement about AI and local news. Local journalism is a vital information source for communities and new sources of data are now available to help people understand their local area, but local news has been in decline. How can we research together with communities how AI could be used to strengthen local journalism and bring communities together in new ways around trusted news? This will require participatory research - and, crucially, exploring where AI should not play a role and why.

• Reaching young audiences in an AI-mediated information ecosystem. Young people are growing up amid rapid technological change and are faced with a vast array of media fighting for their attention. As the UK voting age is lowered to 16 and they have to navigate a world of misinformation and disinformation and AI slop, trusted news and information become ever more crucial. How can we research with young people the diversity of their news interests, priorities and needs as AI-mediation intensifies and information-seeking habits change? This will require longitudinal work and devising ways to meaningfully incorporate these insights into newsroom practice.

So, what’s next?

Public service media has a unique position to take the lead in answering these questions and convening others - including publishers, policymakers, and people - to ensure AI works in the public interest.

Focused research can provide the insights public service media need to make sure they are at the front and centre of the construction of the next era of journalism, taking a value-driven approach to AI integration in the newsroom and playing a distinctive role in bolstering information integrity in a society transformed by AI.

RIC will pursue these priorities by working collaboratively with the BBC and wider news industry, policymakers, regulators, scholars, technology companies, civil society and the public.