No attacker required: the leak that was just a link
The published findings describe no intrusion to reconstruct, and the regulator’s holding does not turn on whether one occurred. A file left a government system, the link that carried it was open, and sensitive personal and medical information about roughly 21,000 special-education students became reachable by whoever held that link. Israel’s privacy regulator called it a severe security event anyway. That holding is why this file belongs on the desk of anyone who litigates data breaches.
What actually happened
Begin with the mechanism, because it is the whole story, and it is almost disappointingly ordinary. Sensitive information was exported from a Ministry system used to place students in educational settings, then became reachable by unauthorized parties through an open link sent to recipients outside the Ministry. The Authority’s own words are that the information leaked and was distributed.
What travelled inside that file is what gives the finding its weight. The data set included names and national ID numbers, age and date of birth, sex, area of residence, educational institution and grade level, religious affiliation, and medical information about the type of each student’s disability. These are children, and several of those fields are on their own enough to identify one.
On 8 October 2026 the Israeli Privacy Protection Authority, the regulator that sits within Israel’s Ministry of Justice while exercising its powers independently, closed an administrative supervision proceeding concerning the Ministry of Education. The Authority had opened the proceeding on its own initiative, using the supervisory powers it held at the time of the event, before Amendment 13 to the Privacy Protection Law entered into force.
A word on the legal furniture, which is unfamiliar outside Israel. Section 17 of the Privacy Protection Law places responsibility for securing a database on the entity that controls it, the database controller. The Privacy Protection Regulations (Data Security), 5777-2017 put operational meat on that duty: a database-definitions document, a written information-security procedure, a risk survey, permission management, and rules for transferring information over a network. Amendment 13, in force since 14 August 2025, added administrative fines to the Authority’s enforcement tools, but it was not in force at the time of this event.
Measured against that list, three deficiencies did the damage. The information moved without appropriate encryption and with no access restriction on the files. It was produced by an employee who did not appear on the system’s list of valid authorization holders. And at the relevant time the Ministry held no database-definitions document as required, had no information-security procedure meeting the Regulations’ requirements, and had not performed an information-security risk survey for the system. (A risk survey that was never carried out is, in fairness, unlikely to have identified the risk that then materialised.)
Two holdings close the file, and both travel well beyond it. First, the Authority determined that the exposure of the information and its use without authorization amount to a severe security event as defined in the Regulations (some English-language sources render the term “serious security incident”), independent of whether the event originated in a cyber attack. Second, the Ministry argued that its ISO/IEC 27001 certification exempted it from complying with the Regulations, and the Authority rejected it, clarifying that reliance on the standard is conditional on meeting the terms in the Authority’s guideline. The Authority concluded that the Ministry breached its duty under section 17 and the Data Security Regulations provisions on the database-definitions document, the information-security procedure, the risk survey, permission management, and transfer of information over a network.
Keep the register right. This is a regulatory finding against a government ministry, not a criminal conviction, and the Authority’s publication announces no fine. Globes reports that none was imposed, noting the event predates Amendment 13.
The artifact: a link’s permission state
Strip the case down to the one object a court would put under the lamp. There is no binary to disassemble. There is a sharing configuration. (Nothing here had to be defeated; the sharing settings had already conceded the point.) The decisive artifact is the permission state of a link, read together with the authorization record of the system the data came from.
That single substitution rewrites the forensic question. In an intrusion case you ask who got in and how. Here you ask what the object’s permission state was, who could reach it, when that state was set or changed, and what the system’s own records prove about the export. Even the missing paperwork is evidence: the absence of the definitions document, the security procedure and the risk survey is part of the proof itself.
What a computer forensics expert witness examines when no attacker is required
This is where a computer forensics expert witness earns its place, because none of the above is self-proving.
Start with the object. What was the link’s permission scope, and what does its change history show about when it was set, and by whom? A link that is open today may have been opened after the export, and that difference decides who was negligent and when.
Then the authorization question. The finding is that an employee outside the valid-authorization list produced the export. That is two separate records: the export event in the system log, and the authorization list as it stood at that exact moment. Authorization lists drift. Proving a mismatch requires the list’s state on the day, not the list as it looks now.
Then scope, which is where claims most often fail quietly. “The information leaked and was distributed” is a conclusion. In a damages proceeding it has to become a measurement: how many retrievals, by how many distinct parties, over what period, and with what confidence, given what the platform logged and how long it kept those logs. A system that keeps thirty days of access records cannot answer a question about month four, and an opinion that pretends otherwise will not survive cross-examination.
Finally, the certification. An ISO 27001 certificate covers a defined scope. (A scope is a boundary, not a force field, however often it is invoked as one.) The real question is what the Statement of Applicability covered, and whether the placement system fell inside the certified scope at all. It is a document-reading exercise with a technical answer, and usually where a certification defence either holds or collapses.
Why this matters beyond one ministry
Almost any organisation holding sensitive data has a version of this file: an export someone made to get work done, passed along through a link because a link is frictionless. Nothing in that sequence requires an attacker.
For litigators the takeaway is narrower, and sharper. If a severe security event does not require an attacker, then a defence built on “we were not hacked” is not a defence at all. And a claim built on “the data leaked” is not yet a quantified claim. Both sides end up needing the same person: someone who can read permission states, authorization records and logs, and say precisely what they establish and where the evidence stops.
The above is general information only and does not constitute legal advice. The facts described are drawn from the publication of the Israeli Privacy Protection Authority listed in the sources.