A One-Week Experiment Giving Passwords to Artificial Intelligence Assistants
During a house move, a writer tested the conveniences and risks of the technology by granting account access to autonomous AI agents from Meta and OpenAI.
Autonomous artificial intelligence agents Muse from Meta and dots from OpenAI were tested for a week by a writer using personal accounts and emails.
The Launch of Autonomous Agents
In September, Meta introduced Muse, which can shop and send emails, while OpenAI rolled out agents called dots for its highest-tier subscribers. These systems are designed to manage users' digital tasks.
Tests During the House Move
During the house moving process, the writer used these agents, nicknamed Mark and Sam, to search for furniture on Facebook Marketplace and manage logistics processes via email.
Encountered Technical Obstacles
While Mark succeeded in digital tasks, Sam faced certain obstacles such as being flagged as a bot by Meta when provided with login credentials.
User Trends and the Future
Within a month of its launch, Muse surpassed three million weekly users, demonstrating the future importance of autonomous agents.
Expert Opinions and Risks
Experts evaluated the psychological, privacy, and economic consequences of outsourcing personal lives to humanized artificial intelligence agents.