The new assistant can draft emails, sell cars and book travel on a user’s behalf, but internal warnings flag potential data‑access risks.
Meta rolled out on Tuesday a long‑touted AI assistant that can autonomously send emails, sell a car and book travel on a person's behalf. The feature leverages large language models integrated with the company's ecosystem, promising to streamline everyday tasks for millions of users.
However, internal engineers raised alarms about the assistant’s broad access to sensitive personal data, warning that mismanagement could expose private information. Critics argue that without stricter safeguards, the technology could become a privacy liability.
The launch comes as the AI arms race intensifies, with competitors like OpenAI and Google pushing specialized models into industry verticals. Investors are watching closely to see whether Meta can balance innovation with responsible data stewardship.