While most employees in Israel now rely on artificial intelligence daily, many organizations still fail to establish clear usage policies.

A high-tech employee in an office

(photo credit: Created by AI)

ByADV. NOA RUBIN

AUGUST 5, 2026 15:35

A Supreme Court petition featured numerous fabricated legal references and rulings. Justice Gila Canfy-Steinitz used the opportunity to stress a wider principle. Even when AI fabricates the sources, the duty to check their validity falls on the professional who submits them. The court made clear that blaming AI for inaccuracies does not shield anyone from professional accountability.

Although the matter concerned an attorney, the ruling carries wider weight. A May 2026 survey by the Bank of Israel and the Israel Democracy Institute showed that 57 percent of Israeli workers now apply AI tools at least weekly. Coders, researchers, attorneys, accountants, marketers, engineers, and HR staff all turn to these systems to generate code, prepare texts, review data, and speed up results. AI has quickly shifted from an emerging gadget to a standard workplace resource.

Who answers when AI produces mistakes?

With wider adoption comes a sharper labor-law question that remains unsettled: when an employee uses AI and an error follows, who is held accountable? Does fault rest only with the worker who should have double-checked the result, or does the employer also share blame for promoting the tools without issuing precise instructions?

Current Israeli statutes offer no direct guidance. The legal duties of care, fidelity, and competence were written for a time when people alone made decisions. Now some tasks run through systems that deliver rapid, convincing answers that may still be wrong. AI can also risk client privacy by exposing personal details. As these tools move into daily professional tasks, the line between human and machine error grows less distinct.

Mistakes can quickly affect employment through lost trust, disciplinary action, or termination. Distinguishing personal fault from an organizational shortcoming in tool deployment is therefore a practical issue that decides who ultimately pays.

Why employers may share accountability

An OECD report noted productivity gains among AI users yet cautioned against unchecked dependence on outputs. Major global firms are introducing internal rules that bar entry of sensitive data into public systems, require human review of results, and state that final responsibility stays with the employee. In Israel most workplaces still lack such policies, leaving staff to adopt the tools on their own.

Microsoft, for instance, requires all staff and contractors to complete security training that covers safe use of generative AI and specifies approved tools and conditions. The firm also supplies an internal portal with model guidelines and task-specific advice. Although these rules do not settle ultimate liability for errors, they show employers recognize the risk and accept a role in prevention through training and direction.

It is hard to expect responsibility to remain solely with the employee. As organizations gain from faster output and time savings, they may also face growing pressure to set usage standards and decide which tasks can safely use AI and which need human oversight.

In time, just as employers must train staff on safety, data protection, and harassment prevention, they may need to provide instruction on responsible AI use. Should that occur, accountability for AI-assisted work will rest with both the employee and the organization.

The author is an attorney at Gideon Rubin and Co. specializing in labor law.