Matthew Watson

NEW: Cognigy is set be acquired for almost $1bn, marking Europe's biggest AI acquisition to date. The German agentic AI startup has been bought out by US customer service giant NiCE — a win for Cognigy investors like Eurazeo and Insight Partners. It's the latest in a string of deals in which American industry scoop up promising European tech companies, from Salesforce's acquisition of Convergence to AMD's $655m purchase of Silo. Read all about it in my latest for Sifted below: https://lnkd.in/e4XYsVCG

91

1 Comment

It may not be next week, or even next month - but this is why the AI bubble will burst. I run through some of the reasons in an extra column for the Comment section at The Telegraph today, but this one is the most important: “The fundamental assumption that huge capital investments are necessary to maintain progress. Billions are being spent on data centres and expensive graphics chips. But in reality, there is no competitive advantage to this, or in the jargon, there’s no ‘moat’ for a business to defend. When Victorian railway companies like the LNER built lines and locomotives at great expense, they could control the trade conducted on them. Later, patents and trade secrets allowed pharmaceutical and software companies to justify huge R&D expenditures, although both had to grapple with piracy. But today every generative AI feature or technique is copied by competitors instantly, within days. Open source models are built specifically to be copied by other people. So a billion dollars of capital expenditure training an AI model suddenly looks like a terrible and stupid waste of capital. Meanwhile the Chinese are focusing on performance, and producing AI models that run well at a hundredth of the expense. We have yet to see this reality reflected in the equity market. Here in the West, we may have backed the wrong kind of AI” Here’s a link to the column: https://lnkd.in/es9SypwU

183

4 Comments

NEW: Preparations for the EU’s code of practice on labelling AI-generated content are moving forward. While no launch date has been set, AI firms, experts and stakeholders will review next week the preliminary results of technical studies meant to shape the code. Summaries of the draft reports on text, audio, and image/video labelling, obtained by MLex, conclude that no single technique satisfies the AI Act’s requirements, recommending multi-layered approaches that combine different methods. The European Commission’s Joint Research Centre has also examined the infrastructure and governance needed to ensure interoperability, raising the prospect that companies such as OpenAI and Google may need to align with common specifications to comply with the AI Act.

126

5 Comments

Scooplet from me: Netflix Co-CEO Ted Sarandos said last week that a wide variety of #creators and #videopodcasters could be a good fit for the streamer. It's further confirmation that the company is looking to broaden the type of entertainment it offers as it chases YouTube. Netflix is taking a concrete step in that direction, quietly putting out feelers to hire an exec to help lead its podcasting efforts. Details in Business Insider: https://lnkd.in/eq_GCTnf

137

8 Comments

This is Tilly Norwood, an AI actor. Talent agencies are in discussions to sign on Tilly. Some of the biggest music labels are about to strike deals with AI companies on licensing AI music. Big-name Hollywood actors and musicians have protested - and have got plenty of coverage - over the disruptions caused by AI. But who's protesting - and getting heard - about the less glamorous jobs being lost to AI? Translators, coders, paralegals, livestreamers ... the list goes on. At Rest of World we've been covering the impact of AI on these less glamorous, but equally important jobs. Read our coverage here: https://lnkd.in/gfH8Y5UC

41

Media folks: it's déjà vu out there as publishers blast Google for helping itself to content without paying—but this time to train its AI. The good news is that the latest version of the content wars may have a happier ending. Cloudflare's CEO (and former lawyer) Matthew Prince says a golden age of content creation is coming, and that Google will follow the example of OpenAI and others and pay publishers to train their AI. “The right way to solve this is not with lawsuits but with royalties,” says Internet OG Bill Gross whose new firm ProRata.ai is building AI tools based around content licensing. This sounds right. Recall how YouTube was piracy central but then a compromise came where creators could either take down or monetize videos. But not everyone buys this! Janice Min of Ankler Media notes Big Tech has burned publishers before, offering content deals and then yanking them away in order to grab ad dollars for themselves (hi there, Facebook!). If you care about this stuff, make time to read this banger of a panel from Brainstorm Tech, or watch the linked video if you prefer your content that way! https://lnkd.in/gjDVP2HD

40

3 Comments

Journalists are overwhelmed - I get 200 to 300 emails a day - so if your subject line doesn’t earn a click, the rest of your pitch might as well not exist. Nailing subject lines is vitally importance, but is an often-overlooked skill in my experience. (Though I've seen some great ones this week as you'll see.) I've put together a few tips below. Tip one is to think like a newsroom, not a product brochure. Then: put the important bit first. Use the inverted pyramid and then repeat it at a sentence level. Also, tell a story, not a slogan. One (anonymised) example of a bad one this week? “Company X and Company Y partner for faster, safer Africa–UK/EU money transfer” is background noise. A good one? “France seeks powers to delist Shein and Temu from Google - comment available” is a story with stakes, actors and consequence. So I clicked - and starred! it in my weird email management system to revisit later. If you want sharper open rates and happier journalists, I run a hands-on workshop for comms teams where we rewrite your actual subject lines, pressure test your hooks, and leave you with a repeatable checklist you can use the same afternoon. It's fast, practical and tailored to the stories you need to land. Get in touch if you want this for your team: https://lnkd.in/eY9s9yJC And if you’re feeling brave, reply with the best - or worst - subject line you’ve seen this week.

43

4 Comments

Did you know that 15% of people aged under 25 now get their news from AI chatbots each week? The latest edition of Reuters Institute for the Study of Journalism's annual Digital News Report shows a rapid shift in the newsroom, where AI is no longer just a buzzword – it’s changing how and where younger audiences access information. But when it comes to AI in the Newsroom, it's not all excitement – there's also a healthy dose of scepticism. The data shows audiences are both intrigued and cautious about AI in journalism, with 66% of respondents expressing interest in at least one AI-driven feature. The strongest appeal is for making news more accessible and usable, not just more personalised. Examples of audience interest: 📝 Summarisation → 27% 🗣️ Translation into multiple languages → 24% ✔️ Better story recommendations → 21% 🤖 Chatbot Q&A → 18% Younger audiences show the greatest appetite – especially for new formats and chat experiences powered by AI. At the same time, many worry about missing important stories if topics are selected on their behalf. In terms of the main perceived pros and cons of the use of AI in journalism: - Positives expected: cheaper, faster, more up-to-date news - Negatives expected: less transparent, less accurate, less trustworthy What I found perhaps most interesting in this section of the report is how attitudes diverge depending on the location of the consumer: 44% of those in India are comfortable with news produced mostly by AI, versus just 11% in the UK. Although I suppose us Brits are pretty famous for our cynicism, so maybe it's not that much of a surprise! 👉 The key takeaway? Audiences want AI to help make journalism clearer, faster and more accessible, but overall trust comes from human oversight and transparency. 💬 How do you feel about the use of AI in journalism? Would you be comfortable getting news directly from an AI chatbot? Or would you worry about what's being left out? (You can read the full report here: https://lnkd.in/gDQ7-ZsS) 

39

5 Comments

"Generic 'what is X?' posts are over. If a default LLM can answer it, why would anyone come to you?" - Kevin Indig ☝️☝️ THIS is what I love about GEO. It's forcing marketers to create original, expert-backed content with an actual point-of-view. Not bland explainers. Not 5000 words of fluff. We're talking about pieces with original insights, quotes, and data. It's almost like journalism skills are still valuable. 😜 Read Kevin's full story on Wix Studio's AI Search Lab: https://lnkd.in/eVyFUEuc

47

5 Comments