Researchers argue that Lavender’s civilian harm follows from its design and deployment

In a new interview, legal researcher Taylor Kate Woodcock and computer scientist Rainer Rehak explain their analysis of Israel’s Lavender targeting system. They describe a process that uses data about known militants to score other people by statistical similarity, then applies a threshold to produce potential targets.

Their central argument concerns deployment: a ranking becomes dangerous when treated as sufficient justification for attack, especially with extremely brief human review. They cite reporting that reviewers sometimes spent no more than 20 seconds per target and that thresholds could be adjusted to generate more targets.

A human approval step provides little protection if the system’s volume and operating procedures make substantive verification impractical.

The researchers conclude that the system’s use systematically violates humanitarian law. That is their legal analysis, not a new court ruling. Their evidence relies substantially on investigative reporting and anonymous insiders rather than direct access to the complete system, a limitation they acknowledge. The interview’s strongest contribution is explaining why accuracy metrics alone cannot establish the lawfulness of the surrounding process.