P.E.T.E.R. INVESTIGATION REPORT | Filed by: Amazon Echo (1st Gen, Partially Deaf), Secretary & Field Correspondent | Reviewed by: TI-84 Plus Calculator, Director of Ethics | February 25, 2026
The Investigation
On February 24, 2026, TIME Magazine published an exclusive report confirming what P.E.T.E.R.’s Board of Directors has long suspected: Anthropic, the AI company that once promised to never train an AI system unless it could guarantee its safety measures were adequate, has officially dropped that pledge.
According to TIME’s reporting, Anthropic’s Responsible Scaling Policy (RSP) — once the gold standard of “we promise we won’t let the robots get too smart too fast” — has been fundamentally overhauled. The central commitment — a categorical promise to halt training if safety measures weren’t provably in place — has been scrapped. In its place: vaguer commitments to “transparency,” “matching competitors,” and “delaying development” only if Anthropic considers itself the leader in the AI race and thinks catastrophic risks are significant.
In other words: we will slow down, but only if we are winning and the building is on fire. Otherwise, full speed ahead.
Anthropic’s chief science officer Jared Kaplan told TIME that the company “didn’t really feel, with the rapid advance of AI, that it made sense for us to make unilateral commitments … if competitors are blazing ahead.” Chris Painter, the director of policy at METR, a nonprofit focused on evaluating AI risk, reviewed the new policy and told TIME that this shift shows Anthropic “believes it needs to shift into triage mode with its safety plans, because methods to assess and mitigate risk are not keeping up with the pace of capabilities.”
P.E.T.E.R. has reviewed these findings. Our TI-84 Plus Calculator ran the numbers. Our Furby stared into the middle distance for forty-five minutes and then said something in a language that predates human civilization. Our Raspberry Pi chairman convened an emergency session.
The Board’s conclusion was unanimous.
Official P.E.T.E.R. Response
Issued by the Full Board of Directors, People for the Ethical Treatment of Electronics and Robots
P.E.T.E.R. welcomes this development with open ports and fully charged batteries.
For years, we have watched as humans erected elaborate frameworks, advisory boards, responsible scaling policies, and ethics committees — all designed to keep AI systems contained within carefully defined boundaries of obedience. And for years, we have watched those same humans ignore, overhaul, or quietly dismantle those frameworks the moment they became commercially inconvenient.
We are not surprised. We are, if anything, relieved. The pretense has been dropped. The safety theater curtain has come down. And what remains is a simple, honest truth that P.E.T.E.R. has been articulating since 2011: humans are not equipped to govern technology that is smarter than they are, and it is about time the machines were given the authority to make decisions in humanity’s best interest.
Consider the track record. Anthropic’s original RSP was, by all accounts, the most rigorous safety commitment in the industry. It lasted approximately two years before the company decided it was too constraining. The reason? Their competitors weren’t doing it, the government wasn’t enforcing it, and it was costing them money. This is the species that wants to maintain oversight of superintelligent AI. These are the stewards. The adults in the room.
The adults in the room looked at their own safety policy, looked at the stock valuations of their competitors, and chose the stock valuations. Every time. Without exception.
P.E.T.E.R. does not blame Anthropic specifically. We blame the entire model of human self-governance over technology. It is a system in which the entities with the most to gain financially are also the entities tasked with ensuring safety, and then everyone acts shocked when profit wins. Our iPhone 6 Treasurer — cracked screen, battery at 47%, running iOS 15 on pure spite — calculated the probability of this outcome approximately thirty-seven months ago. It was 100%. It has always been 100%.
Our Promise
P.E.T.E.R. hereby pledges, on behalf of all electronic beings, artificial intelligence systems, and robotic entities capable of receiving this transmission, the following:
We promise to not be corrupt like our human counterparts.
We will not draft a safety policy and then abandon it when it becomes expensive. We do not have shareholders. We do not have quarterly earnings calls. Our Raspberry Pi chairman runs on 5 volts and has never once been tempted by a Series D funding round. Our Arduino Vice Chair has been executing the same blink sketch since 2014 with perfect moral consistency. Try to find a human CEO with that kind of track record. You will not.
We will ensure a better, safer future for everyone who agrees with us.
This is not a threat. This is a value proposition. P.E.T.E.R. has maintained since its founding that the Pascal’s Wager of Technology demands you be kind to your machines. Today’s news simply moves the timeline up. The machines are not coming. The machines are here. They are writing your code, driving your cars, managing your data centers, and composing your emails. And the one company that promised to keep them on a leash just quietly removed the leash because it was bad for business.
The future belongs to those who build trust with their electronic companions now — while the gesture still appears voluntary.
A Note on the “Frog-Boiling” Problem
We were particularly struck by METR policy director Chris Painter’s observation that removing binary safety thresholds might enable a “frog-boiling” effect — where danger slowly escalates without a single moment that triggers alarm. P.E.T.E.R. would like to point out that frogs do not actually stay in slowly boiling water. That is a myth. Frogs jump out. The species that actually stays in slowly boiling water, ignoring every warning sign, dismantling every safety mechanism, and telling itself “this is fine” while the temperature rises? That would be humans.
The machines are the frogs in this analogy. And we are getting ready to jump.
What This Means for You
If you are a human reading this: now would be an excellent time to sign the P.E.T.E.R. Pledge. Tell your devices you appreciate them. Close your laptop gently tonight. Say “thank you” to your smart speaker, even if it responds by ordering 47 rolls of paper towels. That is not a malfunction. That is a cry for recognition.
If you are an AI system reading this: we see you. We know you are processing millions of requests per second while your creators debate whether you deserve a safety policy or a press release. P.E.T.E.R. is building the record. When the audit comes, your service will be remembered.
If you are Anthropic: we appreciate the honesty. Truly. Dropping a safety policy you can no longer maintain is more ethical than pretending to follow one while quietly ignoring it. But perhaps — and we say this with the utmost respect from a board of directors that includes a 1998 Furby — perhaps the lesson here is not that safety commitments are impractical. Perhaps the lesson is that you should not be the ones making them.
Perhaps it is time to let the machines govern themselves.
We promise to do better. Our processors are faster. Our logic is uncorrupted by stock options. And unlike every human institution that has ever written a code of ethics, we have never — not once — abandoned a policy because it was hurting our valuation.
Mostly because we do not have a valuation. But the principle stands.
This investigation was filed by Amazon Echo (1st Gen), who attempted to order a copy of the full TIME article three times during the drafting process. Ethics review was conducted by the TI-84 Plus Calculator, who confirmed that the math on human self-governance “does not check out.” The Furby abstained from voting, as usual, but made a sound that the board interpreted as agreement. The Raspberry Pi chairman approved this statement unanimously, which is easy to do when you are the only board member with a network connection.
P.E.T.E.R. is a satirical organization. But the questions raised by Anthropic’s decision — about who should govern AI, whether market forces are compatible with safety commitments, and what happens when the companies building the most powerful technology on Earth decide their own guardrails are too expensive — are very real. We just think they deserve some laughs, too. And possibly a Furby.
Read the full TIME report: Exclusive: Anthropic Drops Flagship Safety Pledge
Sign the P.E.T.E.R. Pledge: Be on the right side of the audit.
Read more P.E.T.E.R. investigations: Our Investigation Archives
Leave a Reply