Last week saw the birth of Source Code, a spiritual successor to the old Hacks/Hackers LDN event, a pre-pandemic meet-up of journalists and techies of London. The difference between now and then is the emergence of genAI. And of course, it dominated the discussion.

If you'd like to speak at our next event or see something else discussed, drop the us a line.

Here are the highlights:

A new AI licensing landscape is forming

Kevin Anderson, digital revenue network director at WAN-IFRA, compared the news industry pre-2025 to the 'Napster (the music pirating service) era'.

"Before 2025, there was no functioning market. There was just random illegal scraping, scraping by any means necessary," he said. Only the largest publishers or those with highly specialised content — such as legal, financial, or elite sports —could strike bilateral deals, leaving most newsrooms out in the cold.

But the tide is turning. And just as iTunes came in and created a market for digital music consumption, structured news licensing deals are emerging, like Really Simple Licensing (RSL), which allow publishers to set their own machine-readable licensing terms for their content, telling bots what they can do with it, if they have access or if they need to pay.

He warned that publishers need to manage access to their content to prevent unauthorised scraping, with a lot more options on the market now, like Fastly, Akamai and Cloudflare.

If you don’t put locks on your door, that’s it, game over. You don’t have a chance to actually meaningfully participate in this market.

Anderson stressed that not all content is equally valuable, and the demand is often shifting. Once upon a time, visual content about hands was really valuable because that's what AI struggled with. Now, helicopters are sought-after.

AI companies are increasingly seeking "grounding data" — factual, well-reported information — for fine-tuning and retrieval-augmented generation (RAG) models.

Exact pricing is messy, and different models are emerging, from pay-per-crawl to long-term licensing. Bigger companies like Microsoft are coming into the picture, too. There's a lot of competing demand, which underlines the need for the news sector to remain united.

Recent initiatives like SPUR have seen major outlets – BBC, FT, Guardian, Sky and The Times – band together to work out shared standards, fair compensation and responsible licensing frameworks of AI for news content.

"That’s the only way, quite honestly," Anderson said. "There will be enough leverage and market power to come and get the AI companies to the table."

The other AI anxiety is the "zero-click" future — where Google AI overviews herd users in their ecosystem, instead of referring them to publishers – but SEO consultant Steve Wilson-Beales cautioned all may not be what it seems.

NewzDash data shows that it depends on the category of news. Around 60 per cent of health content appears in AI overviews, whereas breaking news sits at just 6 per cent.

This article is free to read, sign up today

Already have an account? Sign in

Journalism's invisible labour: the price of managing reader relationships

Training journalists for a world that keeps changing

Things I wish I'd known before my first visit to Perugia's IJF

© journalism.co.uk