Newsrooms have been using artificial intelligence for years to do tasks like automation, transcription or content personalisation. But it is the event of generative AI like ChatGPT, that reignited the conversation around opportunities, risks and ethics.
Generative AI will disrupt the way journalism is produced. Rather than wishing it away, journalists need to better understand how they can use this tech to their advantage and counter the threats it poses.
To help your newsroom prepare, FT Strategies, a consulting arm of the Financial Times, held a webinar last week. Journalism.co.uk brings you the highlights from George Montagu, senior manager and head of insights at FT Strategies, who spoke at the event.
The problem with intellectual property
Large language models (LLMs), like ChatGPT or Bard, are heavily dependent on quality input, especially when accuracy matters. In other words, it is the case of 'rubbish in, rubbish out', when it comes to training these models.
Read more: Eight tasks ChatGPT can do for journalists
But here is the problem: these models are trained on hundreds of thousands of pieces of content from publishers like the NYT, the Guardian or the Times. That also applies to content that is behind the paywall.
Publishers have not received any acknowledgement or compensation for the use of their content, despite investing hefty sums in its production.
If you want to see the glass half full, this is an opportunity for media organisations to highlight the value they are creating and try to find ways to claim some sort of payment from the creators of generative AI models.
For those who see the glass half empty, publishers need to urgently redefine how they protect their intellectual property and deal with this situation that will not go away. Digital Content Next principles are a good way to start.
Improve user experience
For years, publishers have been transitioning from simple providers of content to platforms that give users some agency in how they consume the news. This is not just the news provider either - Spotify, for instance, has launched a personalisation feature AI DJ that takes the user from simply listening to songs to actively using a tool.
Publishers may eventually see their content being separated from the format, which will be decided by the user. For instance, they could decide the tone of an article, its length or complexity.
But this is not necessarily good news. Generative AI could lead to even more users getting their news through side doors instead of going directly to the publishers. This could happen through summaries by ChatGPT or Bard, and if that happens, fewer direct users will inevitably affect publishers' bottom line and ad revenue.
To avoid this, news providers must prioritise direct relationships with their audience and make sure they become a destination for their users, instead of just content producers.
Listen: Podcast with Mattia Peretti, former manager of JournalismAI, on responsible use of AI
Premium on trust
There have been some high-profile cases of errors caused by generative AI, aptly called "hallucinations", where the tool makes something up and a news website publishes it without checking.
Fears that generative AI will flood the internet with low-quality content are founded and it is already happening. Although depressing, this is also an opportunity for news brands to become a destination of quality, trusted content that audiences can turn to to make sense of the world.
There is no hiding from the fact that the role of the news media is changing from helping people stay up to date with what is happening in the world, to providing trusted, human-verified high-quality content that helps people make decisions in their daily lives.
But this torrent of AI-generated rubbish will very likely drive down quality and further damage audience engagement. So we need to reconsider what is really valuable to our readers and viewers and double down on quality, trusted, original content that will help differentiate us from AI-generated content.
You can watch the full webinar recording here.
Free daily newsletter
If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).
Related articles
- Humans vs machines: the publishing world's tech tango
- News media must change tactics when fighting false information to protect democracy
- How The Economist reached young audiences through new formats and brand marketing
- What the Financial Times learned from experimenting with AI
- How the BBC is using artificial intelligence