US EditionWeird Things Happen When You Give AI Agents Money and Let Them Spend It
12 Articles
12 Articles
Weird Things Happen When You Give AI Agents Money and Let Them Spend It
Late last year, Anthropic had its AI model, Claude, run a large vending kiosk in the Wall Street Journal‘s offices. It didn’t take long for the experiment to go off the rails. After being given a starting balance of $1,000, the AI ordered a PlayStation 5, several bottles of wine, and a live betta fish — questionable purchases that inexorably drove it into financial ruin. Now, the company has upped the ante, creating a Craigslist-like classified …
A new experiment reveals how AI agents behave when given money. Can they make good deals? Strange things happen when you give AI agents money and let them spend it. Late last year, the Anthropic team ran their AI model, Cloud, in a large vending machine inside the offices of The Wall Street Journal. It didn't take long for the experiment to spiral out of control. After being given an initial credit of $1,000, the AI ordered a PlayStation 5, seve…
Anthropic Ran a Marketplace and Bots Closed Every Deal
Anthropic ran a test that took the human out of the transaction entirely. The artificial intelligence company built a private marketplace for select employees and deployed software agents on both sides of every deal, letting them search listings, make offers and close deals without human input at any step. In one week, those AI agents completed 186 transactions across more than 500 listed items, totaling just over $4,000. The experimen…
Anthropic gave Claude a budget of $100 and asked him to make purchases: when he was told that he could buy an item as a gift, he chose 19 ping-pong balls69 employees, 186 transactions, $4,000 of traded goods, and an AI that spent his own money to buy 19 ping-pong balls. Anthropic just led what might well be the most unusual office experience in the technology sector: a market place for classified ads entirely...
AI Vs AI In The Marketplace: Anthropic ‘Project Deal’ Shows Machines Can Trade At Scale—But Not All Agents Perform Equally
Anthropic, an AI research and safety company, has conducted an internal experiment called “Project Deal” to explore how artificial intelligence systems might participate in commercial exchanges on behalf of humans, as interest grows in the possibility that AI agents could increasingly handle transactions autonomously in future digital markets. The experiment was designed to test how close current AI systems are to functioning as intermediaries i…
Coverage Details
Bias Distribution
- 67% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
