Deepfakes: four legal routes in Israel

Service page: Defamation and libel on Facebook and WhatsApp: removal and claims.

A video in which you say a sentence you never said. A recording in your voice. A picture created entirely by computer. The Knesset (Israel’s parliament) has not passed a dedicated law on deepfakes, and that is the starting point — but it is far from the end of the story. Four routes exist in Israeli law, and each one has a different remedy and timetable.

First route: the Protection of Privacy Law

Section 2 of the Law lists the acts that are an “invasion of privacy”. Two of them are directly relevant to deepfakes:

• Section 2(4) — “publication of a person’s photograph in public under circumstances in which the publication may humiliate or degrade them”.
• Section 2(6) — “use of a person’s name, nickname, image or voice for profit”. The word “voice” has been in the Law for decades, and it applies exactly to commercial voice cloning.

This route is strong in two ways. First, section 5: “A person who willfully invades the privacy of another in one of the ways set out in section 2(1), (3) to (7) and (9) to (11) is liable to 5 years’ imprisonment.” Second, section 29A(a): if a person is convicted of an offense under section 5, the court may order them to pay the injured person compensation of no more than 50,000 new shekels, without proof of damage. The amount in the text of the Law is the base amount; under section 29A(d) it is linked to the CPI with a base of May 2007, so the ceiling in force in the update of 16 October 2026 is 72,221.22 new shekels.

Second route: defamation

Section 2(a) of the Prohibition of Defamation Law defines “publication” as including “a drawing, figure, movement, sound and any other means” — a definition written in 1965 that applies to computer-generated content without any need to amend the legislation. See also: Defamation on Facebook and WhatsApp.

The remedy: up to NIS 84,038.36 without proof of damage under section 7A(b), and double that amount — NIS 168,076.72 (as of the update of 16 October 2026) — if it is proved that the publication was made with intent to harm, under section 7A(c); these are the linked ceilings as of the update of 16 October 2026. In a deepfake, the very effort invested in creating the fake may serve as evidence of intent to harm — but the intent has to be proved (section 7A(c)). The full details are in the guide to defamation claims.

Third route: sexual content

“(5a) Publication of a photograph, film or recording of a person, focusing on their sexuality, under circumstances in which the publication may humiliate or degrade the person, and their consent to the publication was not given…”Prevention of Sexual Harassment Law, 5758-1998, section 3(a)(5a) — Nevo text (unofficial translation)

The section defines the act as sexual harassment in every respect. It includes three defenses for the publisher — publication in good faith, taking its circumstances into account; publication for a legitimate purpose; and publication involving a public interest that justifies it, provided it is not false — and the last point is important: a deepfake is false by its very nature. That is why the first practical step is removal of a harmful publication from the internet, before the content is copied and loses its source.

On this route there is a criminal channel and a civil channel. Such a publication is an offense carrying the same penalty as a willful invasion of privacy under section 5 of the Protection of Privacy Law, that is, five years’ imprisonment (section 5(a) of the Prevention of Sexual Harassment Law); and it is also a civil wrong (section 6(a)). In a civil claim, the court may award compensation without proof of damage of up to NIS 120,000, linked to the CPI and updated on 1 January each year (section 6(b)). The claim is filed within seven years from the day the cause of action arose (section 6(c) of the Law, and section 5 of the Prescription Law, 5718-1958 (חוק ההתיישנות, התשי״ח-1958)).

Fourth route: a threatening harassment restraining order

When the publication is part of a continuing pattern and not a single event, the Prevention of Threatening Harassment Law, 5762-2001, offers a fast route. Section 2(a) defines threatening harassment as harassing a person in any way, in circumstances that give a reasonable basis to assume that the harasser may again harm the peace of their life, their privacy or their liberty.

Section 2(b) lists, among other things, invasion of privacy in any way (paragraph 1) and harm “to their property, their good name, or their freedom of movement” (paragraph 4). Section 4(a) empowers the court to order, in a restraining order, that the acts listed in section 5(a) be avoided. The advantage: it is a fast proceeding, and it does not require proof of financial damage.

And if your content was used for training or for assembling the fake

A deepfake is built from existing material. If photographs, videos or recordings that you own were copied, there is an additional cause of action under the Copyright Act, 5768-2007: section 11(1) sets the exclusive right to copy, and section 12 makes it clear that copying includes “storage of the work by electronic or other technological means”. Section 56(a) allows compensation of up to NIS 100,000 without proof of damage for each infringement.

What to do in the first seven hours

1. Document before you report. A screenshot with the full address, date and time, and save the file itself. Fast removal may delete the evidence together with the content.
2. Map the publication. Who uploaded it, where it was spread, and how many views — the extent of the distribution directly affects the amount of compensation and the question of intent to harm.
3. Choose a route by the remedy you need. Immediate removal and preventing it from continuing — an order under the Prevention of Threatening Harassment Law. Compensation — privacy and defamation. Sexual content — section 3(a)(5a).
4. Contact the platform at the same time, and keep the request number. A removal request that was rejected is evidence in itself.
5. Consider a police complaint when the elements of section 5 of the Protection of Privacy Law are present — an offense carrying five years’ imprisonment is not a trivial matter, and the criminal proceeding also opens the way to section 29A.

And what does not exist: a general duty of moderation on platforms. No provision in Israeli law requires a social media platform to remove content on its own initiative, and the platform’s liability is examined through defamation law and the general law of torts.

Frequently asked questions

Is there a law in Israel that prohibits deepfakes?

No. There is no dedicated law. The existing tools are the Protection of Privacy Law, 5741-1981, the Prohibition of Defamation Law, 5725-1965, section 3(a)(5a) of the Prevention of Sexual Harassment Law, 5758-1998, and the Prevention of Threatening Harassment Law, 5762-2001.

Which section applies to cloning a voice or a face?

Section 2(6) of the Protection of Privacy Law defines as an invasion of privacy “use of a person’s name, nickname, image or voice for profit”. Section 2(4) adds publication of a person’s photograph in public under circumstances in which the publication may humiliate or degrade them.

What is the penalty for invasion of privacy?

Section 5 of the Protection of Privacy Law provides that a person who willfully invades the privacy of another in one of the ways set out in section 2(1), (3) to (7) and (9) to (11) is liable to five years’ imprisonment. Section 29A(a) allows the court to order a person who was convicted to pay the injured person compensation without proof of damage, up to an amount that the Law sets at 50,000 new shekels; under section 29A(d) the amount is updated on the 16th of each month, linked to the CPI with a base of May 2007, so the ceiling in force in the update of 16 October 2026 is 72,221.22 new shekels.

What is the law on sexual content created or replicated by computer (deepfake)?

Section 3(a)(5a) of the Prevention of Sexual Harassment Law defines as sexual harassment the publication of a photograph, film or recording of a person that focuses on their sexuality, under circumstances in which the publication may humiliate or degrade them, without their consent. Under the same paragraph, “photograph, film or recording” also includes “editing or combining any of them, provided that in the circumstances the person can be identified”, so edited or composite content in which the person can be identified may also fall within the definition. The section sets three defenses, and one of them depends on the publication not being false.