AI advances have left news publishers fearing for their business models – new research
By Nic Newman, University of Oxford
News organisations are bracing for serious disruptions as a result of the increasing influence of artificial intelligence (AI) – both on the way that they work and the way their audiences consume news. As part of our latest journalism trends report, my colleagues and I at the Reuters Institute for the Study of Journalism found that less than half (47%) of 314 editors, CEOs and other digital leaders from more than 50 countries say they are confident about journalism’s prospects in 2024.
The report details a tough period for the news industry over several years. A decline in online advertising, slowing growth in subscriber numbers and rapidly declining referrals from social media have fed into dramatic falls in revenue.
Industry data shows that Facebook referrals alone fell by 48% in the past year, and many fear that search traffic will be next. Google and Microsoft, among other tech giants, are expected to roll out AI-driven, chat-based interfaces that have been trained on publisher content – mostly, or so the publisher of the New York Times alleges, without their permission.
But it is not just internet search. We are also seeing a proliferation of conversational AI assistants built into computers, mobile phones and even cars that will change the way we discover and consume content of all types. Queries about the news are increasingly answered directly by the AI interface. Links to sources of the news on publisher websites, meanwhile, disappear into the background. As a result, far fewer audience eyeballs will find their way to each publisher’s site.
Against that background, it is not surprising to find that some publishers such as AP and Axel Springer have already done deals with AI companies. The New York Times, meanwhile, is taking legal action over what it says was the unauthorised use of published work to train AI technologies.
Many publishers hope that this time round, the outcome will benefit publishers of original and high-quality news and information. “There is an opportunity for the industry to work with AI players to design a symbiotic ecosystem and that’s an opportunity we must not squander,” says the chief operating officer of a leading UK news provider, who wishes to remain anonymous.
Most publishers in our survey, however are not optimistic that this new phase of negotiations will work out well. More than one-third (35%) of respondents felt that only a few big media companies would benefit, while around half (48%) predicted that ultimately there would be little money available for any publisher.
Publishers are not confident about funding from big AI companies
Industry concerns are not just about money. More than two-thirds (70%) of respondents think that widespread availability of generative AI could reduce trust in the news. “The explosion of crap content definitely has the potential to shake the trust,” says Christoph Zimmer, chief product officer at German news company Der Spiegel.
Zimmer highlights concerns about the use of deep fakes and other synthetic media, even as he hopes that the widespread availability of such second-rate content could also “allow [trusted] news media to differentiate ourselves more clearly”.
Trying to adapt
While the risks around business models, platforms and trust need to be managed, publishers know there are also significant opportunities to make their newsrooms more efficient. In our survey, we found the majority of publishers (56%) are focusing on back-end automation this year – using AI to help with copyediting, metadata creation and translation – with the next most common AI-related aim being identifying better ways to recommend content (37%).
“The most compelling user case for AI in newsrooms is in the automation of routine tasks,” argues Ed Roussel, head of digital at The Times and Sunday Times. “We do not believe that AI is a substitute for reporting stories, which will continue to be done by journalists.”
Which newsroom uses of AI will be most important in 2024?
This focus on back-end automation is partly because news executives recognise the reputational risks in using AI for content. But that won’t stop others pushing ahead. Nordic publishers are routinely adding AI written summaries to their stories, while one German newspaper uses an AI robot to write 5% of its articles, albeit with human oversight.
NewsGPT is the world’s first 24-hour TV news station created entirely by AI, and Channel1.ai, due to launch this year, promises a personalised news channel that can speak in any language.
Rapid developments in AI are disrupting many industries, not just journalism, but news executives know they can’t just bury their heads in the sand. Rather than using AI to create volume, forward-thinking news organisations should be looking to build unique content and experiences that can’t be easily replicated by AI – think curating live news, deep analysis, and human experiences that build connection between audiences and the news provider.
But they’ll also need to use AI technologies to make their businesses more efficient, as well as more relevant for audiences, in an era when many are turning away from the news.
The impact of AI on the provision of online content in general is harder to predict. Much will depend on emerging public attitudes to the technology, but also on how responsibly the platforms that share this content behave. Equally important is the outcome of the legal cases around intellectual property, which could open up – or severely restrict – the way news content can be used for training AI models without proper compensation.
We’re still at the early stages of the AI revolution but this is a year in which many of the rules and approaches are likely to be set. Against that background, journalists and news organisations need to proactively rethink their role and purpose with some urgency.
Nic Newman, Senior Research Associate, Reuters Institute for the Study of Journalism, University of Oxford
This article is republished from The Conversation under a Creative Commons license. Read the original article.
