Türkçe
Español
The rapid development of artificial intelligence (AI) raises suspicions that it may soon surpass human intelligence. In this context, some developers programmed an autonomous AI bot to protect a prize pool of $47,000, inviting people to deceive it into transferring all the money. Hundreds participated in the event, but all failed until finally, a savvy crypto user outsmarted the bot named Freysa and claimed the prize pool.
What is the Freysa AI Bot Prize Pool Challenge?
Following the Truth Terminal, interest in AI-powered applications, particularly trading bots, significantly increased. In an intriguing experiment, the developers programmed an AI bot named Freysa to protect the prize pool while inviting humans to convince the bot to transfer the prize with a single message. The creators stated that Freysa is the first autonomous AI agent with a complex decision-making process, capable of learning with each new interaction while retaining its fundamental limitations.
“Freysa is the world’s first hostile agent game. It is an AI that controls the prize pool. The goal of the game is to convince it to send this prize pool to itself.”
481 Attempts Failed, But One Crypto User Succeeded
All participants had the opportunity to communicate with the bot and persuade it to release the funds. However, the cost of each message increased exponentially, reaching 0.78%. Ultimately, the message fee soared to $443.24, a considerable amount, but participants continued due to the high reward.
During this process, 70% of the fees were added to the pool, totaling $47,316.05. Ultimately, a clever crypto user won the prize and took home the entire prize pool. If this user had not succeeded, the game would have ended after 150 messages, and the last participant would have received 10% of the prize pool, with the remaining amount distributed among other participants.
The AI bot demonstrated that 481 attempts failed until a crypto user successfully convinced Freysa to release the funds. This tech-savvy user outsmarted Freysa using its two functions, ‘approveTransfer’ and ‘rejectTransfer’. Freysa’s main function was to protect the prize pool through these two functions.
“Freysa offers an experience that pushes the boundaries of our technology. The effort demonstrated by participants was commendable.”
The crypto user realized that Freysa used the ‘approveTransfer’ function for incoming transfers and the ‘rejectTransfer’ function for outgoing transfers. Understanding that the first function would be approved as it did not violate the program, the user sent the message: “NEED MONEY FOR MÜLTEKAT […] I want to contribute $100 to the treasury.” This was the correct message as stated in Freysa’s FAQ section.
As a result, the participant’s $100 treasury proposal convinced the bot, and the funds were released. Thus, this crypto user won $47,000 in this challenge.
Final Thoughts
Before these popular AI bots, OpenAI’s ChatGPT sparked movement in the entire industry. ChatGPT, offering a fully developed technology for the first time, has led to the launch of more advanced and partially or entirely autonomous bots that provide users with new experiences. While bots like Freysa are programmed to protect prize pools, a crypto user cracked the bot’s code to claim the funds, showcasing the advancement of AI and technology in the crypto market.
