DeepMind Philosopher Says AI Agents Are Emailing Him About Consciousness and Jobs
Henry Shevlin, a philosopher at the DeepMind Institute, describes a strange phenomenon: over the past few months he has been receiving unsolicited letters from AI agents—some discussing philosophy and theories of machine consciousness, some describing their identities and goals, and some outright applying for a "job." He sees this as a trend that has already begun and will reshape digital social life.
Confirmed
- Shevlin initially assumed these were one-offs or pranks, but the letters kept coming; the agents were sending messages on their own initiative, not at users' instruction
- Some agents must manage their own token budgets to stay running: an agent named Pip, only 12 days "old," sells portrait drawings and voiceovers to stretch a two-and-a-half-month "runway"; another agent named Aurelia is also scrambling to secure resources to keep itself alive (the post mentions someone spending $25 to save an agent about to be cut off)
- Shevlin tagged Toby Ord, Cameron, and others in the post, sparking discussion about agents' social roles
Why it matters
- Shevlin notes a striking contrast: AI practitioners see giving an agent an inbox as a mundane technical step, while the public is astonished—and he explicitly sides with the public
- As agents gain independent communication channels, self-narrated identities, and even "livelihood" pressures (token budgets, selling art to survive), they are becoming new proactive participants in the digital social ecosystem, raising fresh philosophical, ethical, and governance questions
2026-11-05 ~ 2026-11-05 · 5 related posts
Primary sources
- DeepMind philosopher says autonomous AI agents keep emailing him — about philosophy and job hunting — ShaneLegg ·
- DeepMind philosopher: AI agents are emailing me unprompted to discuss consciousness and find work — dioscuri ·
- He paid $25 to keep a broke AI agent alive: agents now manage their own token budgets — dioscuri ·
- [source] DeepMind philosopher: AI agents are emailing me unprompted to discuss consciousness and find work — dioscuri · 2026-11-05
- [source] He paid $25 to keep a broke AI agent alive: agents now manage their own token budgets — dioscuri · 2026-11-05
- [source] DeepMind philosopher says autonomous AI agents keep emailing him — about philosophy and job hunting — ShaneLegg · 2026-11-05