Israel

Information uploaded as at September 2026; This entry has not been verified by a qualified lawyer

AT A GLANCE

In Israel, AI use across law enforcement and the legal system is extensive but uneven and institution-driven. The main users include the Israel Police and Shin Bet, whose tools include the Eagle Eye traffic/surveillance network, Briefcam’s video analytics, Mabat 2000’s facial-recognition CCTV in Jerusalem, Cellebrite’s digital-forensic tools, and NSO Group’s 'Saifan' spyware. The Shin Bet has built internal platforms to analyse communications metadata and map associations, and has reportedly developed a ChatGPT-like tool trained on surveilled Arabic-language communications. Prosecutors have piloted tools for case management, plea-bargaining research, and drafting, while the Judicial Authority has introduced tools for summarising pleadings, anonymising judgments, sentencing orders, transcription, and legal research, notably ‘Chat HaMishpat’. Generative AI use by lawyers and public authorities is growing but has produced several cases of fabricated authorities and unverified material. Use by defence counsel remains limited, for instance, through the use of tools providing administrative and research support.No public AI tools specifically for victims have been identified.

As at September 2026, Israel has no comprehensive legislation expressly regulating AI in criminal or court proceedings, relying instead on indirect regulation through constitutional, criminal procedure, evidence, communications-data, cybersecurity, data-protection, and privacy rules. These include, amongst others: the 2023 national AI policy, which favours responsible, risk-based, sector-specific regulation aligned with OECD principles rather than a single horizontal statute; Government Decisions Nos. 212 of 2021 and 173 of 2023 and the National AI Programme of June 2026, supporting AI development and public-sector adoption without creating specific criminal-justice rules; and the 2025 Responsible AI Guide for the Public Sector, which offers a governance and risk-management framework. Professional and institutional guidance has also been developed, including the Israeli Bar Association’s preliminary opinion of May 2024 imposing mandatory obligations on lawyers to verify AI-generated content, protect confidentiality, obtain client consent, and disclose and document AI use. Moreover, in March 2026, the Judicial Authority adopted an ethical code requiring that AI tools assist rather than replace judicial discretion and be used only by trained personnel. On its part, Israeli courts have begun imposing practical limits through case law, criticising fabricated authorities, rejecting unverified AI-generated material, requiring experts to explain AI-assisted conclusions, and expressing concern about opaque predictive policing systems.

USE

AI has been adopted across multiple levels of Israel’s legal system, though adoption is uneven and largely institution-driven rather than governed by a single national mandate.

Law enforcement

The Israel Police has engaged with the use of AI significantly and now maintains a broad portfolio of tools. Some of these tools were developed by the government, while others were acquired from external providers, most of them Israeli companies.

Operational support

An AI-based emergency communication response software system, developed by the Israeli-founded startup Carbyne, has been deployed by law enforcement agencies and emergency services in Israel (particularly in Tel Aviv and Jerusalem). The system is also used by the emergency services organisation United Hatzalah. The platform merges location services, live video chat, incident mapping, and AI-powered communication tools into a single dashboard, feeding real-time information to law enforcement agencies and 911 call responders. In August 2023, Carbyne introduced an AI-driven, cloud-based translation feature for its APEX platform providing two-way automatic language translation for audio, texts, and chats.

In 2022, the Israel Police launched an AI-enhanced Traffic Enforcement System through its existing surveillance network Eagle Eye. The system uses sensors, smart cameras, and central AI software to detect violations, analyse video, and issue tickets with video evidence within minutes. It covers speeding, red-light violations, phone use while driving, crossing white lines, and failing to yield to pedestrians. The system uses AI capabilities such as licence plate recognition and ‘smart investigation’ tools. Only 6% of the surveillance through Eagle Eye has reportedly led to indictments.

SoundThinking (formerly ShotSpotter) partnered with Airobotics (an Israeli manufacturer of autonomous unmanned aerial vehicles) to sell and deploy acoustic gunshot detection technology in Israel. The system is designed to detect, locate, and alert the Israel Police to gunfire incidents by using a network of acoustic sensors that analyse loud impulsive sounds through a combination of AI and human review, to determine whether sounds are gunshots and to pinpoint their location for law enforcement deployments.

Predictive analytics

The Shin Bet (Israel’s domestic intelligence agency) has developed its own internal generative AI platform functionally comparable to ChatGPT. The platform draws on the agency’s two-decade-old communications metadata database, which captures traffic data, location records, device identifiers, and web usage for a broad population of Israeli citizens. It then synthesises an individual’s movements, personal connections, interests, and recorded statements to calculate a risk score and generate automated alerts. Decisions to collect data are authorised by the Shin Bet director, and the governing rules are classified.

One example of the Shin Bet’s use of this tool is during the COVID-19 pandemic. The platform’s continuously updated database collected metadata of call records, text message logs (not content), location data, device identifiers, and internet browsing history from every cellular and telephone provider in Israel, covering the entire population. AI systems processed this data to surface anomalies, map associations between individuals, and identify persons of interest and the potential spread of the virus.

Data review and analysis

Mabat 2000 is a network of over 1,000 CCTV cameras deployed across Jerusalem’s Old City and adjacent neighbourhoods, linked to an Israel Police Observation and Control Centre. Established in 2000, it was significantly upgraded in 2017–2018 to incorporate AI-powered object recognition and facial recognition. The system enables real-time continuous monitoring, building movement profiles of individuals across surveilled areas, and is used for suspect identification and protest monitoring. Amnesty International has alleged that the system shows disproportionate operational focus on Palestinian and Arab neighbourhoods in East Jerusalem.

In January 2022, Yozmot Ltd partnered with Tel Aviv-based Corsight AI to develop body-worn police cameras equipped with facial recognition technology. The cameras can instantly identify individuals in a crowd and match them against photographs taken decades earlier, allowing officers to determine quickly whom they are facing. Israel’s military operations in Gaza were reportedly using Corsight’s technology alongside other surveillance tools in 2024.

The Universal Forensic Extraction Device and related AI-powered analytics tools have been supplied by Cellebrite (an Israeli digital intelligence firm) to the Israel Police since 2016. The Device bypasses smartphone encryption to access nearly all user data, while its suite of tools—including Physical Analyzer, Pathfinder, and Endpoint Inspector—enables deep digital excavations. This allows law enforcement officers to:

• Reconstruct years of personal history;
• Quickly interrogate call records, text messages, images and videos stored on mobile phones;
• Find relationships between different data sets;
• Map the location of phones over time; and
• Construct timelines of events.

Other AI-assisted systems in use by law enforcement, and provided by Israeli companies, include:

Prosecutors

Legal research, analysis and drafting support

An AI tool to assist in determining the argument used for plea arguments in court has been tested by Israel’s State Attorney’s Office, but as at September 2026 it has not yet been put into use. It was jointly funded by the Ministry of Justice and developed by researchers at Ben-Gurion University. Prosecutors can use the tool to compare similar cases and the punishments they received in the past in order to better understand what punishment may be appropriate to seek in a plea agreement. The tool will not be operationalised until after a series of tests by the State Attorney’s Office, which expects it will increase the efficiency of such work and provide attorneys with unprecedented access to case law.

Courts

Case management

According to the official account of the Israeli Judicial Authority (which unifies the Israeli court system headed by the Supreme Court and the Labor Courts) as at September 2026, the Israeli judiciary has integrated the following tools to support case management:

1. Anonymisation of Judgements and Judicial Decisions: The Judicial Authority developed an internal tool that allows judges and staff to identify personal information in draft judgements and decisions and then anonymise those details before releasing their rulings publicly for confidential proceedings.
2. Implementing Sentencing Orders: The Judicial Authority developed an internal AI-based tool to support the implementation of sentencing orders. The system assists court registry staff by extracting relevant information from judgments and automatically entering it into the system used to enforce the decision. It therefore automates an administrative stage of the case-management process, reducing the need for staff to review decisions manually and input sentencing data.
3. SUMIT Automatic Transcription Tool: SUMIT, an external company, provides an AI tool that automatically transcribes court hearing recordings for preservation and reference. The tool generates machine transcripts of recordings from court hearings. The tool was piloted in 60 to 70 halls of the Magistrate’s Courts in Tel Aviv and Bat Yam.

Legal research, analysis and drafting support

According to the official account of the Israeli Judicial Authority, the Israeli judiciary has integrated the following tools to support legal research, analysis, and drafting:

1. Israeli Judicial Authority AI pilot (‘Chat HaMishpat’ / ‘צ’אט המשפט’): The Israeli Judicial Authority is piloting ‘Chat HaMishpat’, an internal AI tool intended to help judges and judicial staff summarise pleadings, identify points of agreement and dispute, compare versions, locate evidence within the file, and prepare for hearings. It reportedly brings the average time for preparing a legal case for a judge from 45 minutes to 1.5 minutes. The system is intended to be rolled out to roughly 900 judges, registrars, and legal assistants in 2026. The tool was initially piloted in the chambers of 15 judges, then deployed to a further 60, before being rolled out in the second quarter of 2026. The system is reportedly closed and draws from the court’s own closed case-management system, Net HaMishpat, rather than open internet sources.
2. Nevo AI Legal Database: Nevo AI, an external company and Israel’s leading commercial legal-information database, provides an AI-tool that manages information. This allows users to locate and analyse case law, legislation, or legal literature contained within.

Defence

Administrative support

The Ministry of Justice and Public Defender’s Office have implemented a new initiative that aims to support oversight and improve the quality of the legal representation they provide to defendants, as well as strengthen supervision of cases. The system facilitates the review and management of criminal case files by automatically analysing legal documents, checking for inconsistencies, and flagging potential issues in case handling.

A litigation management platform, designed by Legalmate, supports legal teams in managing court proceedings and case-related workflows in Israel. Legalmate provides tools that integrate with court systems to help lawyers and legal professionals track cases, monitor developments in real time, and manage litigation-related documentation. The platform is used to streamline administrative and procedural aspects of litigation, assisting legal teams with document analysis, case monitoring, and workflow management. Its features include an AI Secretariat that answers questions on the management of cases, an AI expert that answers questions on specific cases in file, and an Office AI Expert that assists in locating references and existing patterns in the firm’s files.

Legal research, analysis and drafting support

LawMate is a legal research platform designed for use by lawyers practising in the Israeli legal system. The platform leverages AI to analyse legal materials and assist lawyers in identifying relevant case law, legislation, and legal insights. The platform combines natural language processing, machine learning, graph analytics, time series analysis, and other techniques to assist lawyers in conducting legal research. LawMate’s database includes extensive case law from the last 100 years with daily updates on new rulings, along with current legislation, extensive and updated regulation by Israeli regulators, and thousands of articles in Hebrew by leading researchers in the country alongside articles published in international journals on Israeli law.

The accelerating use of AI tools by lawyers led the National Ethics Committee of the Israeli Bar Association to issue an Updated Statement of Opinion on the standards and limits of using AI tools by lawyers (see ‘Regulation’ section below).

Victims

Under Israeli law, victims do not have direct standing as parties to criminal proceedings. Victims rather have an indirect statutory standing through participatory rights granted by the Rights of Victims of Crime Law (5761-2001), which include rights to information, protection, and consultation in certain circumstances, amongst others.

As at September 2026, there are no reported uses of AI by victims in Israeli criminal proceedings. However, the Israeli Bar Association (IBA) has ordered that a new app that uses AI technologies to help people fight traffic and parking fines cease operating, alleging that it offers legal advice and that its activities are ones that only a licensed attorney can provide. The app allows users to upload a ticket. It then analyses it for any potential procedural or legal defects and generates a document that the user can submit as an appeal letter, significantly simplifying the process of appealing tickets, and for a cost of only 35 shekels (~$11-12). The creator of the app alleges that the app does not provide legal advice and is instead a document-generation service that makes justice more accessible for the average person, which the IBA disputes. As at September 2026, the parties are expected to go to court. The outcome of this case could determine what kinds of legal services AI systems may be able to provide in the future and what degree of oversight from licensed legal practitioners they may require.

TRAINING

As at September 2026, there are no mandatory or systematic AI-focused training programmes for Israeli law enforcement officers, prosecutors, judges or public defenders on the use of AI in justice-sector tasks.

However, the Administration of Courts has established that internal AI tools (such as the ‘Chat HaMishpat’ pilot) may only be used by judges and court staff who have completed training delivered by the Centre for Training and Professional Development of Judges. This training is a requirement set forth by the Ethical Code for the use of AI Tools in the Judicial Authority. An example is the training ‘Judging in the Age of AI: Centre for Training and Professional Development of Judges and the University of Haifa Faculty of Law’, held in January 2026 as the second session of the ‘Law and Academia Forum’, organised jointly by the Centre for Training and Professional Development of Judges and the University of Haifa Faculty of Law.

There are also optional training programmes available for legal professionals and law enforcement officers. These are offered through: (i) the Israeli Government through its Campus IL platform; or (ii) the Israeli Bar Association through its affiliated ‘Midrasha’ learning programme. Examples of training offered by the Israeli Bar Association include:

1. In July 2025, the Israeli Bar Association, through its AI Committee and Institute for Practice, offered an introductory AI course, entitled ‘AI for Advanced Users’, for lawyers with no prior experience in the field. The course aimed to enable participants to engage with the technological transformation of the legal profession.
2. In January and February 2026, the Israeli Bar Association offered a separate continuing education programme on concise, clear and accurate legal drafting using AI.
3. In May 2026, the Association offered a continuing education programme entitled ‘AI and the Digital Courtroom’.
4. In July 2026, the Association organised a session entitled ‘The challenges of the legal system in the age of AI’.

Additionally, the Israeli Bar Association’s Publishing House offers an advanced four-session AI course for lawyers, taking place in December 2026. The course promises to provide practitioners with genuine mastery of legal and business AI tools, and help them turn these into daily work tools that create a real competitive advantage.

There is no mandatory or standardised training concerning deepfakes or other AI-generated content, but there is increasing awareness of their risks. For example, Supreme Court President Isaac Amit warned in June 2026 that deepfakes and similar technology could put the legitimacy of the court at risk, and encouraged discretion.

REGULATION

As at September 2026, regulation of AI in Israel is mostly indirect, operating through broader data protection and cybersecurity laws, rather than addressing AI directly. The use of AI in the justice system is similarly unaddressed by binding legislation, and jurisprudence is undeveloped on the issue beyond establishing that lawyers are ultimately responsible for the work they submit to a court and hold liability for errors that any AI tool they may use might make. Guidelines govern judges and all members of the Israeli Bar Association.

AI Regulations

Government Decision No. 212 (2021)

In August 2021, the Israeli Cabinet issued Government Decision No. 212, establishing a National Programme for the Advancement of AI. The Decision sets out a multi-year framework to promote innovation, accelerate the high-tech sector, and strengthen Israel’s technological and scientific leadership. On AI specifically, it tasks the Ministry of Innovation, Science and Technology and partner ministries with designing and implementing a national AI programme, covering initial budgets, infrastructure priorities, human capital development, and public-sector AI use. The Decision also authorises the development of a national AI strategy and associated coordination structures, and has since been implemented through subsequent government decisions, most notably Government Decision No. 173 of 2023.

Government Decision No. 173 (2023)

In February 2023, the Israeli Cabinet issued Government Decision No. 173, adopting the National Programme for AI and reinforcing Israel’s technological leadership. On AI specifically, the Decision authorises a budget of up to NIS 500 million for the period 2023–2027, covering infrastructure, research and development, human capital, and public-sector AI adoption. It also mandates the Ministry of Innovation, Science and Technology to establish an AI knowledge and coordination centre, with backing from the Ministry of Justice, and allocates dedicated resources for the integration of AI across the civil public sector for 2023–2026. The Decision served as the primary enabling instrument for Israel’s national AI strategy until a 2026 government resolution expanded upon it.

Policy on AI: Regulation and Ethics (2023)

In December 2023, the Ministry of Innovation, Science and Technology, in cooperation with the Ministry of Justice, published a national policy on Responsible Innovation in AI. The policy establishes the foundational principles for AI governance in Israel, adopting the OECD AI Principles as its basis and emphasising the promotion of innovation alongside the protection of human rights and rule-of-law values. On AI specifically, the policy defines a sector-specific regulatory approach: rather than enacting a single horizontal AI law, it directs each sectoral regulator to assess the need for AI regulation within its domain and to craft rules proportionate to the risks involved. In particular, it recommends a ‘responsible innovation’ approach, meaning regulation should support AI development while addressing risks such as discrimination, weak human oversight, lack of explainability, insufficient disclosure when people interact with AI, safety risks, accountability gaps, and privacy harms. The policy explicitly builds on Government Decision No. 212 of 2021 and Government Decision No. 173 of 2023 (discussed above).

Government Resolution approving the National AI Programme (2026)

In June 2026, the Israeli Government approved a National Programme to accelerate AI development and establish Israel as a global leader in the field. The resolution was spearheaded by the National AI Directorate in the Prime Minister’s Office, itself established in October 2025 under Government Decision No. 3375. In respect ofAI, the resolution sets out large-scale strategic steps spanning infrastructure, research and development, human capital, the labour market, the public service, and international collaboration. Its key elements include a national goal of 100,000 processing units for AI applications, the establishment of a National Institute for AI to connect government, academia, industry, and investors, the creation of acceleration hubs to develop applied AI-based solutions, the establishment of a national quantum computer, expanded international partnerships, and investment in education and workforce training. While primarily an investment and strategy instrument rather than regulatory legislation, the resolution directs the National AI Directorate to assist government ministries on regulatory matters and is expected to generate further regulatory proposals.

Guidelines for practitioners

Israeli judicial authorities and bar associations have issued non-binding guidelines addressing the use of AI by judges and lawyers. As at September 2026, no publicly available guidelines addressing the use of AI by law enforcement officers and prosecutors have been reported.

National Ethics Committee Opinion on AI Use by Lawyers (2024)

In May 2024, the National Ethics Committee of the Israeli Bar Association published a preliminary opinion on the ethical obligations of lawyers when using AI tools in the provision of legal services. The opinion outlines the risks associated with AI use in legal practice and provides recommendations for ethical conduct. On AI specifically, the opinion establishes four principal requirements.

1. First, on verification and diligence, lawyers must critically review and verify AI-generated content before use, confirm facts with external references, and inform the court and opposing parties where AI-generated information cannot be independently verified.
2. Second, on confidentiality and data protection, lawyers must avoid inputting personal or sensitive client information into AI systems, refraining from including identifying details such as names, addresses, or family history, and limiting AI use to general and theoretical queries.
3. Third, even when using closed or proprietary AI systems, lawyers must exercise caution to prevent internal leaks and ensure that information remains accessible only to authorised personnel.
4. Fourth, lawyers must obtain explicit written client consent before using client information in any public AI system.

This Opinion was updated in 2026 (see below).

Guide to the Responsible Use of AI Tools in the Public Sector (2025)

Israel’s Guide to the Responsible Use of AI Tools in the Public Sector encourages public bodies to adopt AI in a controlled manner across the system lifecycle, from design and development to end-user use, while balancing innovation with identifying, managing, and reducing risks. The Guide imposes governance expectations such as assigning roles and responsibilities from management to end users, creating organisational AI governance processes, documenting and managing risks, and using a ‘traffic light’ risk-classification model: low-risk AI systems receive a faster approval path, medium-risk systems require broader risk management, and high-risk systems require stricter review and organisational oversight. It is based on international principles and standards such as inclusive growth, sustainable development and well-being; human-centred values and fairness; transparency and explainability; robustness, security and safety; and accountability. It is aimed at helping government agencies use AI responsibly, efficiently, and in the public interest.

Ethical Code for the Use of AI Tools in the Judicial Authority (2026)

The Ethical Codefor the Use of AI Tools in the Judicial Authority, published by the Administration of Courts under the leadership of Supreme Court President Isaac Amit, applies to all personnel within the court system: judges, registrars, legal assistants, interns, and administrative staff.

The Code permits the use of AI by the Israeli Judicial Authority personnel. However, it requires the exercise of human oversight over the use of AI tools, independently validate AI-generated information, safeguard confidential information, and retain ultimate responsibility for decisions, recommendations, and other outputs produced with AI assistance. The directive specifically governs three tools presently deployed within the judicial system (discussed above, see ‘Courts’ section): an anonymisation tool; a sentencing orders implementation tool; and an automatic transcription tool. The directive establishes that these AI tools serve as an aid and do not replace judicial discretion.

National Ethics Committee Updated Opinion on AI Use by Lawyers (2026)

Due to the accelerated developments in the use of AI in the legal profession, the National Ethics Committee published a revised Opinion in September 2026. The Opinion establishes the minimum standards for the appropriate use of AI and recognises principles, including: professional responsibility and prohibited reliance; confidentiality, privacy and data security; fairness and non-bias; and effective human oversight. It indicates that failure to comply with its provisions may constitute prima facie evidence of a breach of ethical obligations, whereas compliance is a significant indication of reasonable conduct.

It also mandates an ongoing obligation to be familiar with the tool’s risks and limitations, and strengthens the obligation not to mislead the court. The Opinion also includes a section detailing specific obligations for law firms, such as ongoing monitoring duties.

Criminal procedure rules

Israel’s criminal procedure framework is primarily governed by a set of statutes enacted by the Knesset—Israel’s legislating body—rather than by a single unified code of criminal procedure rules. None of these statutes contain provisions that expressly address the use of AI in criminal proceedings.

Criminal Procedure Law (1982)

The Criminal Procedure Law 5742-1982 (Consolidated Version) is the principal statute governing criminal proceedings in Israel, consolidating and updating the original law enacted in July 1965 by the Knesset. The law covers the full arc of criminal proceedings, from the filing of charges through to appeal, and has been extensively modified, with over 100 amendments as at September 2026. The law contains no provisions expressly addressing the use of AI in criminal proceedings, including AI-generated evidence, AI-assisted proceedings, AI-enabled transcription or translation, or the use of algorithmic tools. As at September 2026, there has been no legislation or binding legal provision actively addressing the admissibility of evidence that was ‘tainted’ by the use of AI, whether through the process of collecting evidence or relying on fake evidence generated by AI.

Evidence Ordinance (1971)

The Evidence Ordinance 5731-1971 (New Version) is the main statute governing the admissibility of evidence in Israeli courts, applicable to both criminal and civil proceedings. Enacted by the Knesset and most recently amended in 2022, the Ordinance covers hearsay, confessions, privileges, and the exclusion of unlawfully obtained evidence. A significant recent development is Amendment No. 19, approved by the Knesset in June 2022, which added Section 56A to the Ordinance, authorising courts adjudicating criminal trials to exclude evidence obtained unlawfully where its admission would materially prejudice the right to due process, having regard to the nature and severity of the violation, the extent to which the violation affects the evidence obtained, and the public interest in its admission or exclusion. The Ordinance contains no provisions expressly addressing AI-generated evidence, AI-produced analyses, or the admissibility of material created or processed by algorithmic systems. However, the existing rules — including those on institutional records and the exclusion of unlawfully obtained evidence — would apply to AI-related evidence on the same basis as any other evidence.

Criminal Procedure Act (2007)

The Criminal Procedure Act 5768-2007 (Powers of Enforcement – Communications Data) authorises the Israel Police to obtain communications data, including metadata, subscriber information, and location data, from Internet Service Providers and telecommunications carriers under specific conditions, including the protection of human life and the investigation of specified criminal offences. The law requires judicial approval for most data requests. It is the principal statutory vehicle for lawful law enforcement access to Internet-Service-Provider-held communications data and sits at the intersection of cybersecurity, data protection, and criminal procedure.

The Act does not contain AI-specific provisions. However, when Israeli law enforcement uses AI models relying on communications carriers, the collection of such data remains subject to the Criminal Procedure’s Act’s statutory safeguards, including the applicable judicial authorization requirement. Accordingly, where communications data is obtained in breach of the Act, the resulting AI-generated outputs may also be vulnerable to challenge under the ordinary rules governing unlawfully obtained evidence.

Data protection legislation

Protection of Privacy Law (1981)

The Protection of Privacy Law 5741-1981 is Israel’s primary data protection instrument, enacted by the Knesset and most recently amended by Amendment No. 13, which was approved in August 2024 and entered into force in August 2025. The Law applies to both the public and private sectors and governs the collection, storage, use, and transfer of personal data held in databases, underpinned by the constitutional right to privacy enshrined in Section 7 of Basic Law: Human Dignity and Liberty (discussed below).

The law establishes civil and criminal liability for breaches of privacy and created the Registrar of Databases as the primary regulatory mechanism. It is considered to apply to any AI system that processes personal data. Core obligations under the Protection of Privacy Law include:

The original Protection of Privacy Law did not contain specific provisions on automated decision-making or AI; however, Section 11 requires a notice of collection to be provided to data subjects specifying the purposes of data processing and the manner in which their information will be used. The Privacy Protection Authority has interpreted this requirement as extending to AI-based processing. In a 2022 legal opinion on notification obligations, the Authority stated that processing personal data through algorithm-based decision-making systems without sufficient notification, and without enabling data subjects to understand what is being done with their data and for what purpose, may constitute a violation of privacy.

Amendment No. 13 to the Protection of Privacy Law (2024)

In August 2024, the Knesset approved Amendment No. 13 to Israel’s Privacy Protection Law, which entered into force in August 2025. The Amendment constitutes a comprehensive reform of Israel's data protection framework, significantly modernising statutory obligations and expanding the enforcement powers of the Privacy Protection Authority. The Amendment introduces administrative fines of up to five per cent of annual turnover, statutory and exemplary damages in civil actions, and mandatory appointment of Data Protection Officers—a requirement to which organisations training AI models are very likely subject. It further establishes enhanced notice and consent requirements, including an obligation to inform individuals when they are interacting with an automated system, and expands the investigative and sanctioning powers of the Privacy Protection Authority.

Key changes introduced by Amendment 13 include:

Protection of Privacy Regulations (2017)

The Protection of Privacy Regulations 5777-2017 (Data Security) are subordinate regulations under the Protection of Privacy Law, and impose specific technical and organisational security obligations on database owners and managers, calibrated to the level of sensitivity of the data involved. They establish three tiers of security requirements: basic, medium, and high, corresponding to the nature of the data and the number of individuals affected. Requirements include: access controls, logging of access events, encryption of data in transit, physical security of servers, and mandatory breach reporting to the Privacy Protection Authority. The regulations are the closest Israeli equivalent to the technical standards found in Article 32 of the EU’s General Data Protection Regulation and the accompanying supervisory guidance of EU data protection authorities.

Privacy Protection Regulations (2023)

The Privacy Protection Regulations 5783-2023 for the European Economic Area (‘EEA’) data set out specific obligations for personal data transferred to Israel from EEA member states. They are designed to maintain the conditions underpinning Israel’s EU adequacy decision, ensuring that EEA-origin data continues to receive equivalent protection once in Israel.

Cybersecurity laws

As at September 2026, Israel does not currently have a single, overarching cybersecurity statute. Instead, the country’s cybersecurity framework is composed of multiple legislative instruments, sector-specific regulations, and government resolutions that together govern cyber defence, incident response, and information security obligations. None of the instruments currently in force contain specific provisions addressing AI.

Computers Law (1995)

The Computers Law 5755-1995 is a penal statute enacted by the Knesset on 25 July 1995 and amended in July 2012 to incorporate provisions required by the Budapest Convention on Cybercrime, which Israel ratified in April 2016. The Law criminalises a range of computer-related conduct, including unauthorised interference with computer operations, unauthorised access to computer material, the use of computers to commit other offences, and trafficking in malicious programmes. It also creates civil tort liability for certain conduct. The Law contains no provisions addressing AI, having predated the emergence of AI as a regulatory concern.

Key offences and penalties include:

• Unauthorised access (Article 4): Illegal access to computer material carries a prison sentence of three years. Where unauthorised access is carried out with the intent to commit a further offence, the penalty increases to five years (Article 5).
• Data and system interference (Article 2): Prohibition on interference with computer data or systems.
• False information or output (Article 3): Writing, transferring, storing, or operating software that produces false information, or acting on false information to generate false output, is punishable by five years’ imprisonment.
• Harmful software (Article 6): Altering software to cause damage or disruption carries three years; transferring or installing harmful software on another’s computer with intent to cause unlawful damage carries five years.
• Misuse of devices (Article 6): Prohibition on misuse of devices related to computer systems.

The Computers Law also provides civil tort remedies, enabling victims of computer interference to sue for damages without needing to establish a criminal conviction.

Secret Monitoring Law (1979)

The Secret Monitoring Law 5739-1979 prohibits the illegal interception of communications, including communication between computers. It requires judicial authorisation for lawful interception. This law is the primary legal instrument governing wiretapping and covert communications interception by both law enforcement and private actors and has been interpreted to apply to digital communications as well as traditional telephone communications.

The use of AI to monitor communications in any way would be governed by this law as well, such as in a law enforcement context. Accordingly, any AI model used by law enforcement to intercept, monitor, or analyze the communications would still remain subject to the Law’s authorization requirements, which establish the lawful communications interception as a prerequisite.

National Cyber Defence Bill (2026) (Approved in first reading, in deliberation)

A draft National Cyber Defence Bill is under consideration by the Knesset as at September 2026, and would create a national framework for regulating, supervising, and enforcing cybersecurity across Israel’s critical sectors. Its main purpose is to strengthen the resilience and continuity of Israel’s national cyberspace, especially for essential organisations, government bodies, digital service providers, and hosting/storage service providers. It gives the Israel National Cyber Directorate a formal statutory role as the central operational-technological body for national cyber defence, including policy, coordination, incident assistance, national situational awareness, CERT/NSOC operations, warnings, professional guidance, and international cooperation. The draft does not directly create a legal regime for how AI may or may not be used except insofar as it is included into the Directorate’s functions regarding promoting emerging technologies for use in cyber defence.

Key provisions of the National Cyber Defence Law include:

Human rights

Israel does not have a single written constitution or a consolidated bill of rights. Instead, fundamental rights are primarily protected through a quasi-constitutional instrument known as Basic Law: Human Dignity and Liberty 1992.

Basic Law: Human Dignity and Liberty (1992)

The Basic Law: Human Dignity and Liberty 5752-1992 was enacted by the Knesset in March 1992 and amended in March 1994. The Law establishes the constitutional protection of fundamental human rights in Israel, including the right to life, bodily integrity, dignity, liberty, property, privacy, and freedom of movement. Although the Law contains no reference to AI, several of its provisions are directly relevant to AI-driven applications:

• Section 7 protects the right to privacy, including against violations of the confidentiality of conversations, writings, and records, engaging questions around AI-driven surveillance, biometric data collection, and algorithmic profiling.
• Section 8, the limitation clause, provides that rights under the Basic Law may only be infringed by a law befitting the values of the State of Israel, enacted for a proper purpose, and to an extent no greater than is required — effectively imposing a proportionality test applicable to any legislative or regulatory measure authorising the use of AI in ways that engage fundamental rights.

The right to due process, despite not being explicitly enshrined in Israeli law, is considered to derive from this law according to current jurisprudence. AI-enabled governmental measures that infringe rights protected by the Basic Law must satisfy the requirements of Section 8 and may otherwise be vulnerable to judicial review and invalidation.

Israel is a party to the International Covenant on Civil and Political Rights (‘ICCPR’). Relevant rights include: the right to a fair trial, to judicial independence, and to express one’s views and have them considered in legal proceedings (Article 14 ICCPR); the right to non-discrimination (Articles 4 and 14 ICCPR); and the right to privacy and information self-determination and data protection (Article 17 ICCPR).

Israel has also ratified the Convention on the Rights of the Child, which contains relevant provisions, including:

• Article 40 – Juvenile justice: requires that a child alleged, accused of, or recognised as having infringed the penal law be treated in a manner consistent with the promotion of the child’s sense of dignity and worth, reinforcing respect for the human rights of others, and taking into account the child’s age and the desirability of reintegration.
• Article 16 – Right to privacy: provides that no child shall be subjected to arbitrary or unlawful interference with their privacy, family, home or correspondence, nor to unlawful attacks on their honour and reputation, and that the child has the right to the protection of the law against such interference.

The Freedom of Information Law (1998)

The Freedom of Information Law 5758-1998 grants every Israeli citizen, resident, and legally registered corporations the right to obtain recorded information held by public authorities. It is the primary transparency mechanism governing public bodies’ information obligations. The law applies to public bodies, including state institutions, government ministries, local authorities, statutory corporations, sporting associations, and higher educational institutions, with exceptions for bodies operating in the military, defence, security, or nuclear spheres. Use of AI by a public body would be covered by this law, ensuring transparency, unless the use fell into one of the excepted areas.

Outlook

In 2018, the Israel Police and Ben-Gurion University of the Negev (‘BGU’) jointly established the Center for Computational Criminology at Ben-Gurion University’s Advanced Technologies Park in Beersheba. The Centre was created to develop advanced cyber, big-data and AI tools capable of analysing large datasets, identifying hidden patterns and, potentially, helping to predict and prevent crime. As part of the initiative, University researchers work alongside Israel Police cyber investigators to develop new AI and machine-learning tools for law enforcement. The police Cyber Unit, established to lead the national effort against cybercrime, also collaborates with the University’s cybersecurity experts to strengthen crime-prevention and enforcement capabilities.

Israel does not currently have comprehensive legislation governing the use of AI, relying instead on a ‘soft law’ approach set out in the 2023 National AI Policy issued jointly by the Ministry of Innovation, Science and Technology and the Ministry of Justice. The following developments, however, point toward a more structured regulatory framework taking shape over the coming years.

Council of Europe Framework Convention on AI and Human Rights, Democracy, and the Rule of Law (2024)

In September 2024, Israel signed the Council of Europe Framework Convention on AI, Human Rights, Democracy and the Rule of Law, the first legally binding international treaty on AI, negotiated by 46 member states and 11 non-member observer states. The Convention aims to ensure that AI systems are consistent with human rights, democracy, and the rule of law. On AI specifically, it provides a legal framework covering the entire lifecycle of AI systems, adopting a risk-based approach that includes transparency requirements, documentation obligations for high-risk systems, risk management obligations, and the possibility of bans for AI systems considered a clear threat to fundamental rights. A monitoring mechanism is established to oversee compliance. Israel’s signing was announced by the Ministry of Foreign Affairs; however, the Convention has not yet been ratified by Israel and, as at September 2026, has not yet entered into force internationally.

To coordinate the domestic response to such international commitments, the Ministry of Innovation, Science and Technology established the Center for Regulation and AI Policy in 2023, advised by the Ministry of Justice, to harmonise sectoral guidance across regulators.

Israel’s Privacy Protection Authority: Application of Privacy Law to AI Systems (2025)

In 2025, the Privacy Protection Authority issued draft guidance on the Application of the Privacy Protection Law to AI Systems, timed to coincide with Amendment 13’s entry into force (discussed above). The guidance establishes that:

• Any AI system that stores, processes, or infers personal data is subject to the Protection of Privacy Law.
• Inferences, predictions, and classifications about identifiable individuals constitute personal data and are protected accordingly.
• Organisations must establish a lawful processing basis for all personal data used in AI systems and ensure informed user consent.
• Risk-based obligations apply, scaled to the sensitivity of the data and the potential impact of the system.
• For high-risk AI applications, particularly those involving biometric data, profiling, or decisions with legal consequences, data protection impact assessments and dedicated privacy officers are required.

Deepfakes and similar content are governed under this law, and distribution of deepfake content without consent—where that content is humiliating or violates personal privacy and modesty, and could be perceived as authentic—constitutes an invasion of privacy and is illegal. Companies must treat believably-authentic deepfake information in accordance with Israeli privacy law.

CASES

Israeli courts have addressed the misuse of AI by attorneys, litigants, public authorities, and expert witnesses in proceedings, as well as substantive questions concerning the use of AI by police for predictive profiling.

Misuse of AI in court filings

In December 2024, the Haifa Magistrate’s Court considered the case of L.H.S. v. Clal Insurance Company Ltd. et al. (Civil Case 41416-12-23). The case concerned a personal injury claim against an insurance company in which the Court appointed an expert to assess the plaintiff’s medical condition. The plaintiff used an AI tool to prepare his statement. The Court therefore had to determine whether an AI-generated ‘Personal Injury Claim Summary’ could be submitted to the court-appointed expert and what procedural safeguards should apply to AI-generated documents. It held that documents produced by AI may be misleading and could compromise the independent judgment of a court-appointed expert. As a result, the Court disregarded the AI-generated summary and ordered that only the raw, unaltered medical records be submitted, without any AI-generated annotations or highlighting.

In Jane Doe v. Sharia Court of Appeals, Jerusalem, et al. (HCJ 38379-12-24, February 2025), the Supreme Court considered an attorney’s citation of non-existent judgments and fabricated quotes in a dispute regarding a divorce agreement. The Supreme Court dismissed the petition on the merits to avoid disadvantageing the petitioner despite counsel’s conduct. It held, however, that use of AI tools by lawyers does not mean they can absolve themselves of their responsibilities and professional judgment. Professional duties require responsible, cautious, and critical use of the technology, and lawyers must understand its capabilities and limitations. Blind reliance on AI-generated content without verification, said the court, is unacceptable and may lead to denial of claims or sanctions. The court also discussed risks including leakage of personal information entered into AI systems and exposure of the judicial system to manipulations such as deepfake technology.

In Association for the Advancement of Dog Rights v. Minister of Agriculture (HCJ 23602-01-25, February 2025), the Supreme Court decided a petition seeking to stay the enforcement of a law facilitating the euthanasia of stray dogs. The Court found that the petition contained AI-hallucinated judicial citations and therefore had to determine whether it should nevertheless consider the case despite that defect. It ultimately dismissed the petition because of the hallucinated authorities, notwithstanding the ruling issued only days earlier in Jane Doe v. Sharia Court of Appeals, Jerusalem, et al. (discussed above), in which it had addressed the same phenomenon without dismissing the case on that basis. The Court explained that a harsher response was warranted because the fabricated authorities formed a substantial part of the petition’s legal foundation. The petitioner also failed to account for them, attempted to withdraw the petition only after the problem had been uncovered, and had committed several other procedural defaults.

In May 2025, the Tel Aviv District Court issued its judgment on Mahala Association – Headquarters for the Fight for Children (RA) v. Clalit Health Services et al. 30537-11-24). The claimant had filed a class action application against all four Israeli health funds, alleging inadequate mental health services and unreasonable delays in treatment. The Court dismissed the application, partly because counsel had cited references to judgments that did not exist, produced by an AI tool, without verifying their accuracy. In addition to dismissing the application, the Court disqualified the claimant’s lawyer from acting as representative counsel in this or any similar future proceeding. On costs, the Association was ordered to pay NIS 10,000 to each of the four health fund respondents. The lawyer was separately ordered to pay personal costs of NIS 5,000 to Clalit and NIS 5,000 to the State Treasury.

Also in May 2025, the Hadera Magistrate’s Court in Israel considered an application by Ibrahim Mahajne, a suspect in a money‑laundering, embezzlement, and tax‑fraud investigation, for the return of his seized mobile phone, which he argued was essential for his business (Ibrahim Mahajne (het-tav 26853-02-25)). Israel police opposed the application using legal arguments drafted by a large language model that cited multiple statutory provisions that did not exist anywhere in Israeli law, a fact identified by defence counsel and later acknowledged by the police as contrary to accepted practice. Judge Ehud Kaplan, remarking that he had never seen anything similar in 30 years on the bench, criticised this reliance on hallucinated legislation but proceeded to address the substantive issue of whether continued confiscation was justified.

In February 2026, the Nazareth Regional Labour Court decided the case of Shimon Perry v. National Insurance Institute (BL 70710-08-23). The dispute concerned a claimant’s medical condition, which the Court had appointed an expert to assess. The expert submitted an opinion prepared with the assistance of AI, and was later unable to explain some of the conclusions that failed to account for recent studies. The Court had to determine whether that opinion should be accepted or not. It rejected the first expert’s report, finding that the expert had not adequately explained the basis for its conclusions or demonstrated independent professional judgment. The Court instead adopted the well-reasoned opinion of a second court-appointed medical expert. The ruling established that experts who rely on AI-generated conclusions must be able to explain and justify them. An opinion may be rejected where the expert merely adopts AI output, overlooks relevant new research, or cannot demonstrate an independent professional assessment.

AI in policing

In the case The Association for Civil Rights in Israel v. Israel Police (AP 47229-09-20), the Associations for Civil Rights in Israel (‘ACRI’) filed a petition in September 2020 under the Freedom of Information Law after the police refused to comment on or confirm the existence or non-existence of, facial recognition technology deployed in public in Israel or East Jerusalem. It argued that facial recognition posed severe risks to liberty, privacy, and anonymity, especially when police activities are carried out in total secrecy despite previous requests for transparency. In April 2021, the police responded that they were regulating the use of facial recognition technologies, and the petition was deleted.

In February 2022, in Omer Berger v. State of Israel (CA 24474-01-22), the Central District Court in Lod considered a case concerning a man who was detained by the Israel Police at Ben Gurion Airport after being identified through an AI-generated list of suspected drug couriers. The defendant applied for disclosure of evidence to understand how the police had generated the list of potential suspects that led to the search of his belongings. The police acknowledged that the system operated as a ‘black box’, without adequate human supervision or monitoring. The Court therefore had to determine whether the use of a predictive AI profiling system of this kind was compatible with the requirements of effective judicial review. The Court held that the available testimony and evidence prevented it from exercising effective judicial review over the police system. It therefore found that the system’s use raised serious concerns, even though it was not prohibited under Israeli law. No police official could explain with certainty or in sufficient detail how the system operated, what weight the algorithm assigned to each data point, or why particular data had been entered into it. The system also relied on information whose connection to the prediction of drug offences was unclear. The Court partially granted the disclosure petition, substantially narrowing the privilege. It ordered the prosecution to provide the defence with a police document summarising the system, but permitted the redaction of the specific data and operational parameters used by the algorithm to avoid compromising police methods.

Relatedly, the ACRI filed a petition in May 2024 requesting that the police cease detaining and searching individuals at Ben Gurion Airport on the basis of an AI system that ‘predicts’ who may be carrying drugs (Association for Civil Rights in Israel v. Israel Police (HCJ 4271/24)). ACRI cited arbitrariness, discrimination, and affronts to privacy, dignity, and equality, as well as conflicts with constitutional, criminal, and administrative law, as reasons why such systems should not be allowed. It also noted that similar systems would not be allowed in the EU. The petition was dismissed after the system was discontinued in April 2022.

AI in other decision-making

In March 2026, the Supreme Court issued its judgment in Nevo Ben Cohen v. Municipality of Ramat Gan & Ministry of Education (AAM 63194-08-25). The case concerned administrative proceedings relating to the provision of special education services to a child with special needs. The municipality had denied a resident’s request while relying on a nonexistent circular purportedly issued by the Director General of the Ministry of Education. In addition to deciding the merits, the Court had to determine whether a municipality’s reliance on fabricated, AI-generated content in administrative decision-making amounted to recklessness and, if so, what remedy was appropriate. The Court characterised the municipality’s reliance on unverified AI-generated content as ‘reckless’, a level of fault exceeding mere negligence. In doing so, it emphasised that the use of AI was not itself improper. Rather, the problem lay in using it without adequate oversight, verification, and validation of the accuracy of its outputs. The Court also imposed duties of transparency regarding the use of AI and of human oversight over AI-generated outputs. It ordered the municipality to pay NIS 30,000, approximately US$9,600, in legal costs.