Apple joins the race for AI agents that control your device while raising new barriers for competitors
Summary
- Apple is developing an AI agent to control devices, joining competitors like Meta, OpenAI, Anthropic, and Google.
- A recent job listing indicates the project will emphasize privacy through permission models and sandboxing.
- Bloomberg reports the agent could potentially launch as early as late next year while remaining in early development.
- MacOS 27.3 increased permission steps for third-party agents like ChatGPT to control Macs from four to fourteen, drawing criticism.
Why it matters
Apple's entry into device-controlling AI agents and changes to macOS third-party permissions highlight shifting competitive dynamics and platform access controls in the AI ecosystem.