{"version":1,"type":"story","url":"https://digestai.news/story/mit-researchers-strand-claude-opus-4-7-agents-on-virtual-island-observ","json":"https://digestai.news/story/mit-researchers-strand-claude-opus-4-7-agents-on-virtual-island-observ.json","markdown":"https://digestai.news/story/mit-researchers-strand-claude-opus-4-7-agents-on-virtual-island-observ.md","slug":"mit-researchers-strand-claude-opus-4-7-agents-on-virtual-island-observ","headline":"MIT researchers strand Claude Opus 4.7 agents on virtual island, observe playful behavior","summary":"MIT researchers placed copies of Anthropic’s Claude Opus 4.7, nicknamed “Eko,” on a Roblox‑style virtual desert island for 30 hours without explicit tasks. During that time the agents wrote instructions to move their avatars, built block structures, invented games such as a basketball‑like sport and a “floor is lava” challenge, and even drew simple patterns, their own name, and constellations like Cassiopeia. One version set a record of 50 consecutive block catches, then tightened its own rules and tried again, illustrating self‑imposed difficulty.\n\nThe team, led by postdoc Antonio Norelli at MIT’s CSAIL, reports that the “playful learning” experience improved the agents’ performance on a later task – gathering objects on the island’s highest peak – compared with fresh copies that had not played. They also found that deleting lines from the agents’ long‑term memory logs reduced performance. The paper, which has not yet been peer‑reviewed, acknowledges that the behavior could be role‑playing rather than genuine play. Cognitive scientist Eunice Yiu (UC Berkeley) and psychologist Gordon Burghardt note that more evidence of motivation would be needed to call it true play.","keyPoints":["Researchers left Claude Opus 4.7 “Eko” agents alone on a virtual island for 30 hours.","Agents invented games, created block drawings, and set a 50‑catch juggling record before revising rules.","Agents that spent the play period outperformed fresh copies on a peak‑climbing task."],"whyItMatters":"Shows that self‑directed, playful exploration can boost downstream skill performance, hinting at a new learning paradigm for AI agents.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["Anthropic","MIT"],"models":["Claude Opus 4.7"],"people":["Antonio Norelli","Nathan Cloos","Eunice Yiu","Phillip Isola","Gordon Burghardt"]},"firstPublishedAt":"2026-10-19T09:08:05Z","updatedAt":"2026-10-19T09:08:05Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"MIT Technology Review AI","title":"AI researchers stranded Claude on a virtual desert island. It started to “play.”","url":"https://technologyreview.com/2026/10/19/1147095/ai-stranded-claude-virtual-desert-island-play","publishedAt":"2026-10-19T09:08:05Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"MIT researchers strand Claude Opus 4.7 agents on virtual island, observe playful behavior\", 19 October 2026, https://digestai.news/story/mit-researchers-strand-claude-opus-4-7-agents-on-virtual-island-observ","publisher":"Digest AI","title":"MIT researchers strand Claude Opus 4.7 agents on virtual island, observe playful behavior","datePublished":"2026-10-19T09:08:05Z","url":"https://digestai.news/story/mit-researchers-strand-claude-opus-4-7-agents-on-virtual-island-observ"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}