As AI systems evolve to become more autonomous, the role of employees is shifting from users to managers of these agents. Companies are currently lacking training programs that focus on how to lead and oversee AI, which could lead to overtrust and mismanagement of these technologies.
Companies have spent the last two years training employees on basic AI usage, primarily focusing on prompting skills.
The latest AI systems are becoming more autonomous, requiring employees to take on roles that involve directing and overseeing these agents.
Gianpaolo Barozzi from Cisco highlights that this shift necessitates new approaches to setting boundaries, trust, and accountability.
Most organizations are not preparing employees for this management role, which could result in excessive trust in AI capabilities.
A study found that 76% of executives view agentic AI as a coworker rather than a tool, indicating a shift in perception.
Employees often approach AI agents through five mental models: tool, intern, service provider, teammate, or expert, each with different expectations and risks.
The effectiveness of AI depends on how it is managed, with the need for clear oversight and defined roles.
Companies must develop 'human-agent fluency' to recognize the appropriate relationship with AI agents and manage them accordingly.
Research indicates that framing AI as an employee rather than a chatbot can improve error detection, emphasizing the importance of language in managing AI relationships.
Ultimately, human accountability remains essential, regardless of how AI is perceived or utilized.