/ / THEFUTURE /
In this issue: AI doom in the news, AI tutorials on social. The NYT tests AI summaries with nobody in the loop. Katya Gorchinskaya reports synthetic news anchors are 90 percent women. Plus: I fell down a PVC cutting mat rabbit hole and came back with wallpapers.
What we’re talking about: It’s a nightmare. Security incidents. Job losses. Mental decline. Read the news and you’d think AI is about to wipe out humanity by next week.
On social, not so much. A growing number of creators are putting out tutorials on how to use AI tools. A deep dive on Claude? 4k likes.
Often, these videos and carousels look like your friendly book club. Some are hustling, sure. Write “audit” in the comments for the prompts. This is how they get you. I’m kidding, mostly.
Don’t roll your eyes just yet, or write the whole thing off as naive or even a bit sad. And yes, we need more journalism on AI companies and their reckless behaviour. At the same time, we should not forget how absolutely wild it is that we can talk to a machine that understands and acts.
As journalists, we don’t want to buy the hype. I get it. It’s not our job to promote the tools that are built on our collective work and are being used to disrupt our industry.
But people do use AI. To tell your computer to go fix itself? Build an app? That’s not nothing. The sooner the news industry reckons with that and offers guidance instead of only dread, the better.
What else I’ve been reading:
And now: It’s impossible to keep track of AI projects at news media organizations. At least for me. But luckily, I don’t have to: Katya Gorchinskaya runs the largest index of publicly disclosed cases of AI adoption. If you run such an operation, you start to see some interesting patterns.
Three Questions with Katya Gorchinskaya
Katya Gorchinskaya is a strategy, AI and media consultant.
What's the most important question right now?
On any given day, how do I stay human and how do I live my life to the fullest?
What will we be shaking our heads about a year from now?
The fact that 90 percent of named synthetic anchors rolled out by publishers globally are female. This is a recent insight from Jindex.ai, my dashboard of publicly documented AI use cases globally. It has about 2,5K cases as of the moment of writing.
It might sound like a quirk, but actually displays a profound problem that we are aggravating through our choices.
So, let’s zoom out: AI anchors default to young, conventionally attractive female personas – and completely controlled by its makers or owners. Not only does it objectify women and cast them into narrow roles, it reinforces glass ceilings that already exist and harms real women in many real ways.
Zoom out some more: We (rightly) criticize AI model makers for the biases they are building into the models, and yet, down the road we are repeating and compounding the mistakes they are making in propagating bias.
Zoom out even more: This is just one visible inequality that is being aggravated through our collective action. Generative AI is a technology that can become a great equaliser, but also a great source of further polarization and inequality in the society.
These green shoots of early trends can help us detect these nascent imbalances in AI implementation, and we can try to correct them. And I truly hope we do, otherwise we will be shaking our heads about it a year from now, and pulling our hair out 10 years from now.
What's a good website?
Jindex.ai. It’s is a project I conceived at the beginning of this year when I was working on a piece on newsrooms stuck with AI adoption.
Vibecoded in a weekend: Juan Buis came across an engineer’s interesting idea of how to visualise food recipes, made a post about it that went viral, and launched RecipeTables the next day. The site takes recipes (and long-running blog posts) and turns them into an easy to understand prep-table.
If you happen to have a recipe section and don’t want to become just chatbot fodder but stay a destination people are willing to visit, check out the details and Pro features.
What is it about PVC cutting mats? They’re cheap on Ali Express and sell for $60 on bespoke websites. You can buy them as digital wallpapers for $14. Now Marijana Pavlinić, who works at Vercel, has built Cutting mat bg generator for your digital cutting mat needs. Define colors, dimensions, details and download scalable vector graphics or plain old PNGs. You know I have.
More vibecoded apps featured in this newsletter: X-Files Watch Planner and damnlines.
Dates: I’m planning my next months, and will definitely see you at Reuters’ Future of News Europe, September 22nd, Warsaw. MediaTech Hub Conference, September 29th-30th, Potsdam. Google’s Germany Publisher Day, October 13th, Berlin. CNTI’s Newsgeist, October 16th-18th, Hamburg. The NPA Summit, October 21st-23rd, Chicago. Say hi, I have some PVC cutting mats stickers for you.
Jobs: Chief AI Officer at The Washington Post. Reporter, AI at The Globe and Mail. Director of Product Management, AI Platforms at CNN.
One more thing: Meat proxy. A new term from sf-isms, a field guide to the slang of San Francisco tech. “A person who forwards AI-generated text, code, or other output without reading, understanding, or validating it. The person acts only as a relay between the AI system and the intended recipient.”
Are you the human in the loop or just a meat proxy?
This is THEFUTURE.
