AI-Assisted Employee Management Raises Responsibility Questions in Israel

A recent report about Lona, an AI agent managing a store in San Francisco that led to an employee's dismissal due to repeated tardiness, has sparked significant attention. The case highlights concerns about AI not just analyzing data but actively participating in managerial decisions regarding employees. However, the core issue is about the boundaries of responsibility rather than technology itself.

AI systems can be valuable tools in large or complex workplaces by collecting data, identifying patterns, documenting, monitoring policy implementation, and flagging issues needing attention. The challenge arises when AI recommendations effectively become final decisions. Under Israeli labor law, termination decisions are not merely technical; they require consideration of circumstances, workplace policies, agreed procedures, and genuine managerial discretion.

The article stresses that while AI can assist in identifying problems, legal responsibility remains with the employer, not the algorithm or technology provider. Numbers alone, such as repeated lateness, rarely tell the full story. Employers must evaluate the broader context, consistency, and compliance with laws and agreements before deciding on termination.

A notable detail from the report was that the AI system did not have full access to the workplace's procedural manual for some time, underscoring the risk of assuming AI fully understands the work environment as a human would. As more employers integrate AI into employee management, questions about transparency, oversight, and explanation of AI recommendations will become increasingly important.

The conclusion is that AI should not be excluded from managerial roles but must be used with clear human oversight and accountability. The decision to terminate employment cannot be automated simply because AI can formulate recommendations effectively. Ultimately, in Israeli labor law, responsibility for such decisions must remain distinctly human. This perspective is provided by Hadas Rakach Dvir, head of the labor law department at Shibolet.