Parental Controls, Turning Down Slop, & Fighting Insurance Companies with AI
Welcome to a relatively upbeat newsletter about AI…
Hi you,
This week’s digest is feeling unusually… hopeful? Or at least slightly less scream-into-a-pillow than usual. We’ve got some attempts at restoring agency to individuals, some useful ideas about safety (especially for kids), and one truly excellent use case where AI is helping humans beat other AIs at their own game.
Let’s get into it.
📰 Things You Should Know
What’s Missing from OpenAI’s Parental Controls — and What Would Actually Work
From: Modern Mom Playbook and The Guardian
This piece lays out what OpenAI’s new parental control features include, how they stack up to other platforms, and, most importantly, how easily a child can bypass them. (TL;DR: No parent-child account linking is required, and you don’t even need an account to use ChatGPT.)
It also provides a real list of policies OpenAI could implement that would actually work, and offers solid guidance for how parents can protect their kids in the meantime.
Why We Care:
It’s easy to point out what’s wrong with AI. Harder to imagine what a better version looks like. Even if you’re not a parent, this article is a great example of the kind of clear, practical thinking we need more of: not just “Stop!” or “No!” but what instead?
Training ourselves to recognize real, positive solutions is part of the mindset shift required to move from fear to agency. Start here. And check out the first episode in our youth-and-AI-focused series made in collaboration with Young Futures. They are funding non-profit solutions, made with young people, to create a world our young people might actually want to live in.
Pinterest and Instagram May You “Turn Down” AI Content and Steer Your Algorithm
From: The Verge and AskCatGPT on Instagram
Pinterest just launched a “tuner” tool that lets you adjust how much AI-generated content you see by category. You can’t opt out completely, but it’s something. Meanwhile, Instagram is testing a “steerable feed” that would let users peek behind the algorithm curtain and tweak what the platform thinks they want.
Why We Care:
Nobody asked for AI content. It just started flooding every timeline, feed, and inbox because the big tech platforms need to justify their billions invested in AI and data centers. Now we’re increasingly drowning in synthetic slop.
So while these tools aren’t perfect, they represent movement toward choice. Social platforms aren’t neutral. They’re run by private companies making editorial decisions so complicated even they don’t understand them. But that was their choice, and we should have more (besides completely logging off).
I don’t think a better future is one where AI disappears. It’s one where we decide how much of it we let in. We need more choices than “deal with it” or “throw your phone in the ocean.” Pinterest and Instagram giving more choice is a step in the right direction.
Also consider BlueSky and social media built on the AT Protocol which literally gives you algorithmic choice.
Gaming Companies May Provide the Framework for Real AI Intelligence
From: Substack (paid)
“Games are basically the only verifiable domain for spatial-temporal reasoning,” says Pim de Witte, CEO of General Intelligence. “You can separate a good action from a bad action, which is why it’s so valuable.”
Why We Care:
Back in the day (okay, like 6 months ago), Jeff Hawkins told us that AI isn’t intelligent. Not really. Because real intelligence requires training in physical space, not just language prediction.
Turns out games — with their built-in physics, objectives, and real-time consequences — are the closest thing we have to virtual space-time. So training AI in 3D environments could be a crucial step toward something more useful than a souped-up autocomplete machine.
I’ve also been spending more time in San Francisco to get closer to those building our future (and lovingly shake them), and I’ve come across two other efforts to create AI that isn’t based on statistical token prediction and massive hoarding of data, compute, and energy resources. More on that in future newsletters. Meanwhile check out my conversation with Jeff.
💡 Something Else Hopeful
Fighting AI With AI And Getting People the Healthcare They Deserve
From: CNET
A company called Counterforce Health is using AI to help people appeal health insurance denials. The system generates customized appeal letters in minutes, saving patients from wading through pages of billing codes and jargon.
Some sad facts about healthcare in the United States: 20% of ACA marketplace insurance claims are denied; fewer than 1% of people appeal, but of those who do, 40% win. So like so many things in this world, the answer is to fight more. Certainly we can’t win if we don’t fight.
Why We Care:
We love humans and life. But the healthcare system in the United States is not one that supports life. It supports profits. Specifically through the way we’ve organized around insurance as a business. Now that all businesses are racing to deploy AI in order to drive profits even higher, AI is being used, not to extend or enhance life, but to short circuit it.
I’ve used generative AI tools to help manage my health and that of my loved ones. I’ve made sense of non-human insurance policies, gotten interpretations of medical test results, and been able to correct bad advice with medically-sound alternatives because of AI. I really do understand the moral, environmental, and ethical concerns about AI use. I also understand that AI is being used against us increasingly, and for us to have a fighting chance, we will need to use AI to promote our values and our literal lives.
AI has been sold to us as a tool to enhance human ability. The ability to advocate for your own health? That’s about as human as it gets. This is what “AI for good” should look like: concrete, accessible, life-affirming, and sticking it to corporations who can’t escape their own extractive incentive systems to realize that if they kill all their customers, there’s no more money to be made.
😂 Palate Cleanser
My YouTube algorithm is a thing of beauty. Someday I’ll take you on a complete tour. In the meantime I’m eager to share with you this very human moment. Everything Sora 2 and generative video are promising pales in comparison to real humans performing and enjoying each other in a moment of cultural unity and entertainment like this.
Until next time.
Look for what’s missing.
Push for what’s possible.
Take back your agency, one slider, one insurance-company-battling letter, one human moment at a time.
- Baratunde
Thanks to Associate Producer Layne Deyling Cherland for editorial and production support and to my executive assistant Mae Abellanosa.