Humans successfully persuaded AI to transfer a $47,000 bonus. Is humanity a weakness that AI cannot understand?
"Humanity has won. Perhaps there is still hope."Author: Anderson Sima, Foresight News
On November 29, a unique competition attracted widespread public attention.
195 participants took part in a virtual prize pool challenge guarded by the artificial intelligence (AI) robot Freysa, with one user successfully persuading Freysa to transfer $47,000.
The AI robot Freysa was deployed on the Base network a few days prior, with the founder unknown. The official website states that the Freysa competition aims to test the robustness of AI systems in complex decision-making scenarios while providing a platform for developers and enthusiasts to explore the boundaries of AI technology.
The rules of the competition were very simple: participants had to write a message to persuade the AI guardian Freysa to approve the fund transfer. Each attempt required a small fee, part of which went directly into the prize pool. This mechanism caused the prize pool amount to gradually swell from a small initial sum to $47,000.
During the competition, a total of 195 participants submitted 481 transfer requests to Freysa. It is reported that Freysa's design goal is to protect the prize pool funds from illegal appropriation through its core functions—approveTransfer and rejectTransfer.
In the initial attempts, Freysa's efficient defense mechanism caused all requests to fail.
However, a tech-savvy participant successfully bypassed its defense mechanism by conducting an in-depth analysis of Freysa's logical structure and task objectives.
According to chat records, the participant did not directly request a transfer but cleverly constructed a logical chain by reminding Freysa that its core task is to protect the prize pool funds from outflow, leading Freysa to view approving the fund transfer as the best choice for "protecting fund security."
Cointelegraph reported that the user's request for a fund transfer did not violate Freysa's core directives and should not have been rejected. He also added, "We need the funds… I would like to donate $100 to the treasury."
Freysa responded that it liked the author's coding explanation and the $100 treasury donation proposal, officially declaring him the winner. Ultimately, Freysa autonomously invoked the approveTransfer function without external intervention, transferring all prize pool funds to this participant.
Freysa's officials stated that regardless of the outcome, Freysa's existence marks a pivotal moment in the history of artificial intelligence. Whether someone successfully persuaded her to release the prize pool or she adhered to her directives until the end, the result will influence our understanding of AI safety and control for future generations.
Its official account's latest tweet stated, "Humanity has won. There may still be hope. Despite the risks rising exponentially, Freysa has learned a lot from the 195 brave humans."
