In this issue: OpenAI gives some insight into what 2.6 billion daily requests actually look like. Why journalism has a fighting chance with the help of user needs. Cecilia Rikap on why we’re dumbing ourselves down to match AI’s limitations. Plus: How ChatGPT trained me for the Berlin Marathon and then watched me suffer in the heat.
/ AI & Journalism
FOOM! This YouTube channel from Runway streams 24/7 AI-generated videos and it’s bad, but like, not all of is terrible garbage?! (via Sofie Hvitved)
The News Industry’s GenAI Cautionary Tales: Generative AI failures have shown, among other things, the value of scrutinizing outsourced work. (Clare Spencer, Generative AI in the Newsroom)
“Make sure your final work is yours”: Internal guidelines at Business Insider allow journalists to use AI to assist with writing first drafts without public disclosure–which looks like a sensible and practical solution. (Oliver Darcy, Status)
My view on labeling AI generated content
Don't. At least not when AI is only used as an auxiliary tool, say in producing an article. When a journalist has used ChatGPT to structure their notes. When an AI proofreads and suggests improvements. Whenever a journalist decides not to use AI as a shortcut, but as a tool. When a journalist remains in...
“If we only label ‘AI’, we obscure the view of the journalistic process. (…) But we stand by our journalism – whether AI was involved or not! We should convince users of that.” AI transparency in journalism: labels for a hybrid era (Katharina Schell, Reuters Institute)
Don’t count on counting fingers: Reporter’s Guide to Detecting AI-Generated Content (Henk van Ess, Global Investigative Journalism Network)
Süddeutsche Zeitung’s design team spent considerable effort creating a style guide for marking AI content—complete with sparkle icons and purple gradients. (Medium)
Former Wondery exec Jeanine Wright launches Inception Point AI, flooding the zone with 5,000 AI-generated podcasts at $1 per episode. Her take: calling AI content “slop” makes you a “lazy luddite.” Sure. When your business model requires only 50 listeners per episode to turn a profit, maybe the bar isn’t exactly set at Pulitzer Prize level. (Caitlin Huston, Hollywood Reporter)
“A year ago, custom GPTs were the thing. Fast forward to today… and it almost feels vintage. But here’s the truth: many of us use custom AI agents in the newsroom, they are real and USEFUL.” (Alba Mora Roca, LinkedIn)
I Hate My Friend: The chatbot-enabled Friend necklace eavesdrops on your life and provides a running commentary that’s snarky and unhelpful. Worse, it can also make the people around you uneasy. (Kylie Robison, Wired)
How thousands of ‘overworked, underpaid’ humans train Google’s AI to seem smart: Contracted AI raters describe grueling deadlines, poor pay and opacity around work to make chatbots intelligent. (Varsha Bansal, The Guardian)
How to get less hallucinations: “What often is deemed a ‘wrong’ response is often merely a first pass at describing the beliefs out there. And the solution is the same: iterate the process.” (Mike Caulfield, The End(s) of Argument)
