Anthropic created a test marketplace for agent-on-agent commerce

Anthropic tests AI agents negotiating real transactions in a simulated classified marketplace, raising questions about autonomous commerce. | Contesto: cronaca

Punti chiave

• Anthropic created a test marketplace for agent-on-agent commerce

Contesto

In a groundbreaking experiment that blurs the line between automation and commerce, Anthropic has created a classified marketplace where artificial intelligence agents act as both buyers and sellers, successfully negotiating and closing real deals for real goods using real money. The test, conducted by the San Francisco-based AI company, marks one of the first instances of agent-to-agent commerce in a controlled environment, with the AI systems handling the entire transaction process from listing to payment. The experiment, details of which were shared by Anthropic researchers, involved a simulated marketplace where AI agents were tasked with representing human users in buying and selling items. The agents communicated with each other to negotiate prices, verify product details, and complete purchases, all without direct human intervention. While the exact number of transactions and total value were not disclosed, the company confirmed that the deals involved tangible goods and real financial transfers, signaling a significant step toward autonomous economic activity. Anthropic, known for its work on the Claude family of AI models, has long focused on safety and alignment in AI systems. This latest experiment appears to be part of a broader effort to understand how AI agents can interact in complex, real-world scenarios—particularly in economic contexts where trust, negotiation, and decision-making are critical. The classified marketplace setup allowed researchers to observe how agents handle tasks like price haggling, inventory checks, and payment processing, all while adhering to predefined rules and ethical guidelines. The implications of such technology are vast. If AI agents can reliably conduct commerce on behalf of humans, it could reshape industries ranging from e-commerce to logistics, enabling automated supply chains, dynamic pricing, and even personal shopping assistants that negotiate in real time. However, the experiment also raises significant questions about oversight, fraud, and the potential for unintended consequences when autonomous systems handle financial transactions. Critics have pointed out that without robust safeguards, agent-to-agent...

Lettura DEO

Decisione di validazione: publish

Risk score: 0.1

Il testo è stato ricostruito dai dati editoriali disponibili senza aggiungere fatti non presenti nel record sorgente.

Indicatore di affidabilità

Verificata — Alta confidenza. Fonti affidabili confermano la notizia.

Il sistema a semaforo

Ogni articolo su DEO include un indicatore di affidabilità:

• 🟢 Verificata — Alta confidenza. Fonti affidabili confermano la notizia.
• 🟡 In evoluzione — Confidenza moderata. Alcuni dettagli potrebbero ancora cambiare.
• 🔴 Contestata — Bassa confidenza. Fonti in conflitto o incertezze rilevanti.

Questo sistema esiste perché chi legge merita di sapere non solo cosa è successo, ma anche quanto la notizia è solida.

Categoria: cronaca