Protester’s health wanes as AI safety demands go unanswered
Guido Reichstadter has now spent 24 consecutive days without solid food, maintaining his vigil outside the San Francisco headquarters of Anthropic. The 56-year-old activist positions himself daily at 500 Howard Street, where he watches food deliveries arrive for employees of the AI company while consuming only electrolytes and vitamins to sustain himself.
His protest, which began on September 1, centers on what he describes as the “great and profound evil” occurring within Anthropic’s offices. Reichstadter leads the group Stop AI and demands that CEO Dario Amodei acknowledge the existential risks posed by advanced artificial intelligence systems. He specifically wants Anthropic to call for an immediate halt to what he characterizes as a reckless race toward superintelligence.
Fellow protesters abandon their strikes
The hunger strike initially had company. Michael Trazzi and Denys Sheremet began parallel protests outside Google DeepMind’s London office, but both have since ended their fasts. Sheremet stopped after 16 days without calories, expressing hope that DeepMind leadership would “make a first step towards de-escalating the race towards extremely dangerous AI.”
Reichstadter, however, continues alone. His persistence highlights what might be seen as a growing divide between AI laboratories focused on scaling their models and activists who view the entire enterprise as an existential gamble. I think this tension has been building for some time now, but perhaps it’s reaching a new level of visibility with actions like this.
Anthropic maintains silence amid growing attention
Local media outlets including the San Francisco Standard and SFGate have documented Reichstadter’s daily presence, yet Anthropic has not publicly commented on the strike. This silence seems notable for a company that markets itself as an industry leader in AI safety. Reichstadter accuses the company of promoting what he calls “the lie that the race to superintelligence can be done ‘safely’ or ‘responsibly.'”
His protest methods aren’t entirely new—earlier this year he chained open doors at OpenAI’s office and staged demonstrations at Google DeepMind. But the current hunger strike represents perhaps his most extreme action to date. The physical toll must be considerable after nearly a month without proper nutrition.
Public response remains uncertain
Hunger strikes operate on a delicate calculus. Their effectiveness depends largely on whether the surrounding community chooses to pay attention. Reichstadter has wagered his health on the chance that people will notice and that the conversation about AI safety will shift as a result.
On platforms like the Effective Altruism Forum, his opening statement was published, calling for an immediate cessation of what he terms “reckless actions” by Anthropic. But whether this will translate into broader public concern or policy changes remains unclear. The situation feels precarious—both for Reichstadter’s wellbeing and for the larger debate about how society should approach advanced AI development.
After 24 days, the outcome still hangs in balance. Will Anthropic engage? Will public pressure mount? Or will this protest end as others have, with little substantive change? Reichstadter continues his vigil, watching food deliveries arrive for others while his own body weakens, waiting for someone in power to acknowledge what he sees as an emergency unfolding in plain sight.
