
While the public debate in Israel is focused on when there will be an "Israeli AI law," the Privacy Protection Authority is already acting — through the existing law. Its approach is simple: an AI system that processes personal information is a database, and the Privacy Protection Law applies to it.
What Has Been Published So Far
In April 2025, the Authority published draft guidance on the application of the Privacy Protection Law to artificial intelligence systems. It is the first document of its kind by an Israeli regulator on the subject. The comment period was later extended, which indicates the volume of comments received.
At the end of December 2025, the Authority published additional guidance on implementing privacy-enhancing technologies (PETs) in AI systems — tools designed to bridge the tension between the need for large volumes of data and the requirements for protecting personal information.
In February 2026, the Authority published its final opinion on consent in privacy protection law — an interpretive document that raises the bar for what counts as informed consent.
Why Amendment 13 Changes the Equation
Until Amendment 13 to Israel's Privacy Protection Law, the Authority's guidance was mainly an interpretive statement. The amendment granted the Authority expanded administrative and criminal enforcement powers, including significant administrative fines. The Authority itself stressed that the guidance will serve as the basis for exercising these powers.
At the same time, Amendment 13 requires many organizations to appoint a Data Protection Officer (DPO). This is not a symbolic position — it is a role with defined responsibility.
The Hard Part: Consent
The Authority holds that using personal information about a person in an artificial intelligence system requires that person's informed consent. It is a requirement that is easy to declare and hard to implement: information collected years ago, for a different purpose, is not necessarily covered by the consent given at the time.
An organization considering feeding historical customer data into an AI system — for training, analysis or even summarization — first needs to ask on the basis of which consent it is doing so.
What This Means for Your Organization
- Map which AI systems are already running on personal information in your organization. In most organizations the answer is surprising — summarization, transcription and support tools came in under the radar.
- Check whether you are required to appoint a Data Protection Officer under Amendment 13. It is an obligation, not a choice, and it depends on the scope and type of information.
- Before any AI project on personal information — establish the basis for consent. It is the easiest question to skip and the most expensive to discover late.
- The PETs guidance is a practical tool: it offers ways to use data without exposing it. It is worth examining before abandoning a project for regulatory reasons.
