In the transaction, two AI agents communicated with one another using cryptocurrency tokens.
The successful conclusion of the first cryptocurrency transaction fully controlled by artificial intelligence (AI) has been announced by Coinbase CEO Brian Armstrong.
In a recent post on X, the crypto boss shared that AI bots executed the transaction without human intervention.
Using cryptocurrency tokens, one AI agent—basically, a bot created to carry out particular tasks—interacted with another AI agent.
AI-to-AI Transaction Intended to Acquire AI Tokens
Acquiring AI tokens—data strings that enable algorithms to learn from and adapt to new information—was the aim of the AI-to-AI transaction.
“They used tokens to buy tokens,” Armstrong wrote.
Armstrong noted that the incapacity of AI agents to conduct transactions is currently one of the main obstacles facing them.
These AI systems struggle with everyday tasks like making travel reservations or overseeing social media activities other than content creation because they are not able to process payments.
Armstrong said, “AI agents can get cryptocurrency wallets, but they can not get bank accounts.” These AI-driven transactions on Coinbase’s Base platform are instantaneous, available everywhere, and cost-free.
“This is a critical step toward enabling AIs to perform valuable tasks. These days, an AI agent cannot complete a task and return in a few days or hours with any meaningful work completed.”
— Brian Armstrong (@brian_armstrong) August 30, 2024
The advancement fits into a larger trend in the cryptocurrency space that gives AI agents transactional capabilities.
This month, the blockchain development company Skyfire unveiled a payment platform that lets AI agents handle and spend money on their own.
In a similar vein, Web3 infrastructure company Biconomy is incorporating AI agents to facilitate on-chain transactions, providing a fresh authorization layer that lets users assign trading tasks to AI.
AI Agents Should Have Crypto Wallets
Large language models (LLMs) are the core technology of AI systems such as OpenAI’s ChatGPT. Armstrong proposed incorporating LLMs into the crypto ecosystem last week.
He made the case that giving AI agents access to cryptocurrency wallets could greatly increase their usefulness and let them engage in economic activity on behalf of people.
A research report published by Coinbase in June of last year stated that there is a significant business opportunity at the nexus of blockchain technology and artificial intelligence.
The biggest cryptocurrency exchange in the US at the time emphasized the advantages of fusing the two technologies to develop fresh approaches to the problems AI was posing to society.
According to David Duong, head of research at Coinbase, “as applications within AI and blockchain mature, the disruptions these technologies represent may lead to areas of collaboration and the emergence of new use cases for crypto to help address specific societal challenges posed by AI.”
The paper listed a number of possible applications that integrate blockchain technology with artificial intelligence.
First, it claimed that decentralized data marketplaces could assist generative AI in satisfying the need for a diverse and validated set of data for model training.
In a similar vein, incentive systems based on tokens may raise the caliber of data obtained from those markets.
Additional use cases include enhancing data authenticity, increasing the auditability of the opaque decision-making process of AI algorithms, and supplying computational power from decentralized networks that employ graphics processing units for AI projects to train their models.