“Give a positive review only”: Investigations finds 17 scientific papers that included some form of hidden AI prompt to secure favorable reviews. (Shogo Sugiyama, Ryosuke Eguchi, Nikkei)

“While journalists perfect their craft for human readers, technology companies are building parallel infrastructure for AI consumption. The risk isn’t replacement, it’s irrelevance.” (Shuwei Fang, Splice)

Responsible Tech Summer Reading List 2025 (All Tech Is Human)

AI-powered news app Particle expands into Long Reads, teaming up with The Atlantic.

You shall not pass… unless you pay: Cloudflare has launched a private beta enabling customers to allow, charge, or block bot-access to their content. For it to work, bot operators such as OpenAI or Google would have to use a new protocol named Web Bot Auth and be willing to pay.

Case Study on iterative prompt evaluation and improvement: A workflow for targeted prompting to refine AI-generated newsletter headlines. (Ashlyn Wang, Generative AI in the Newsroom)

AI for Newsrooms: Sergei Yakupov has collected 93 initiatives and 58 resources. (Previously shared: Kalle Pirhonen’s spreadsheet with 90+ tools and products from around the world.)

Illustrator Christoph Niemann confronts his fears about AI art. (New York Times Magazine)

BBC launches generative AI features: ‘At a Glance’ summaries and ‘Style Assist’ to adapt and reformat stories so that they match house-style.

A new report is called “No Turning Back: AI’s Growing Role in News”, but shows that newsrooms use AI cautiously, hoping for modest gains rather than disruptive transformation. (Felix Simon, Aspen Digital)

A plea for publishers to ditch their walled-garden AI approaches and embrace emerging open standards like Model Context Protocol. (Ben Werdmuller, werd.io)

Time spent on ChatGPT is approaching that spent on social platforms. (Coatue’s 2025 Market Report via Lucy Küng)

Initial US rulings, as expected, give AI training the green light. Some lawsuits against AI companies have been dismissed, with courts distinguishing between AI training (legal) and traditional piracy (still illegal). (Florent Daudens, HuggingFace)

Google AI Mode Live In US: Tests Deep Search, Live Search, Personalization, Custom Charts, Shopping & Agentic. And no, Search Console data will not include breakdowns for AI Mode and AI Overviews. (Barry Schwartz, Search Engine Roundtable)

So… is ChatGPT making us dumb? A viral MIT study, what it actually found, and the delicious irony of how we consume research. (Steffi Kieffer, Substack)