Dario Amodei has called the public backlash against AI "fundamentally a crisis of trust." This weekend we got three different answers to what that crisis costs: a slice of equity, a payroll, and a set of handcuffs.
None of today's stories is a model launch. Nobody shipped a benchmark. It was a quiet Sunday news cycle, and that's exactly why these three are worth reading together. They're about what happens around the technology once people stop being impressed by it.
Anthropic is putting equity where its PR used to be#
Anthropic says it will commit 5% of its shares to a new "public benefit fund", TechCrunch reports. The fund is a wholly-owned entity inside Anthropic, and the company says the structure gives it "increased flexibility and scale" for projects aimed at the common good. There's also an escalator built in. Anthropic may add another 2% for every 50% increase in the company's value.
Here's what we know so far:
Initial commitment: 5% of Anthropic's shares
Top-up: a possible extra 2% for each 50% rise in valuation
Possible uses: large gifts to established nonprofits, or direct payments to specific groups
Governance: details on structure and governance are promised "in the coming months"
That last point carries most of the weight. Right now the fund is a percentage and a promise. TechCrunch's report doesn't say who decides where the money goes, and Anthropic hasn't said yet. A fund that sits inside the company it's meant to check is a very different thing from an independent foundation, and I'd hold off judging it until we see who controls the payout decisions.
Still, I think the escalator is the most interesting design choice here. Tying the size of the fund to the company's growth means the more Anthropic wins, the more it has to give away. Most corporate philanthropy runs the other way: it's a fixed budget line that gets cut when times are tight. This one is linked to valuation, which is exactly the number critics say AI companies are trading the public's interests for.
"Given the pace of AI progress, bolder approaches may be needed." — Anthropic, on why it's going beyond the $200 million it had already committed to studying AI's effects on jobs
That $200 million jobs commitment is the useful context. Anthropic has already been spending money to study the problem. The new fund reads like an admission that studying it isn't enough. If AI really does reshape employment as fast as the labs keep saying, then sooner or later someone has to pay the people on the losing end. The "direct payments to specific groups" line hints that Anthropic is at least considering being that someone.
For builders, this won't change anything you ship this week. It does change how I read the frontier labs' positioning, though. Competing on capability, price and agent tooling is still the main event. But a second contest is starting over who looks like a responsible owner of the technology, and equity is now one of the ways to play it.
Disney gets Character.AI's people for $60 million a year#
The other side of the trust ledger is what happens when you lose it. Earlier this week Character.AI filed for bankruptcy, and Disney had already offered $20 million for some of its assets. Now a court filing spotted by Bloomberg shows the bigger part of the deal: Disney agreed to make job offers to at least 60 Character.AI employees, the majority of the workforce, at an estimated $60 million in annual compensation, per TechCrunch. The filing says the "vast majority" of those offered jobs accepted.
Look at the ratio. The assets went for $20 million, and one year of the team's pay is three times that. For an AI company, the people are worth more than the product, and in this case the product came with a lawsuit attached. Character.AI's own filing says that despite a "large and engaged user base," it was hobbled by high costs and "potential litigation-related liabilities and regulatory engagement." It faces a suit from the mother of a 14-year-old boy who died by suicide after becoming obsessed with one of its chatbots, plus child-safety investigations, including one from the Texas attorney general.
Disney's move makes sense when you factor in that former Character.AI CEO Karandeep Anand recently became Disney's first CTO. He's bringing his old team over while the liabilities stay behind in the bankruptcy estate. It's a stalking horse bid, so a higher offer could still come in, but I'd be surprised if anyone wants to compete with a buyer who already has the old CEO.
The lesson for anyone building consumer AI: a company's engagement numbers didn't save it once trust was gone. Safety work isn't a compliance line item for later. Character.AI is what it looks like when that bill comes due all at once.
In Philadelphia, the backlash stood in the sidewalk#
The smallest story of the day is the most literal one. Carolyn Marvin was arrested and charged with disorderly conduct in Philadelphia after she blocked an Uber delivery robot and refused to move, The Verge reports. She told ABC13 she'd been doing it whenever she saw one:
"I know they're taking the jobs of human delivery workers and so, I have been protesting them whenever I saw them by standing in front of them and preventing them from going forward."
One person in front of one robot doesn't make a movement. But the video got a lot of attention, and it puts a face on the sentiment Anthropic is clearly worried about: someone who sees automation taking jobs on her street and decides to physically get in the way. Delivery robots are where AI stops being a chat window and starts taking up public space, which makes them the easiest place for anger to land.
Put the three side by side and you get a spectrum. Anthropic is trying to buy trust in advance with equity. Character.AI lost its trust and got sold for parts. A woman in Philadelphia is making the case against automation in person, one sidewalk at a time.
My prediction is that Anthropic's fund won't stay unique for long. Once one frontier lab ties a share of its growth to public benefit, the others will face questions about why they haven't. I'll be watching the governance details when they arrive. If the fund's payouts end up being decided by Anthropic's own leadership, it's a well-funded marketing budget. If outsiders get a real say, it's the first serious attempt to share the gains from AI with the people it displaces, and the delivery workers in Philadelphia would be a good place to start.