AI Hack Watch - timeline and dataset of hacking incidents(JSON/RSS)
- Source
- aihackwatch.com
- Author
- Satinel
- Date
A new project called AI Hack Watch published a sourced, methodology-driven public timeline of real incidents where AI materially participated in hacking or cyber operations, available as both JSON and RSS feeds. The dataset spans lab-disclosed autonomous red-team findings from frontier AI companies through criminal use of coding agents, including reported cases of tools like Cursor being used for live network intrusions. Rather than aggregating headlines or speculation, each entry is sourced to primary disclosures, giving security teams and researchers a running, queryable reference for how AI capabilities are actually showing up in offensive use, as opposed to how they're discussed. For engineers building or securing AI-adjacent systems, a structured feed like this is more useful than one-off news coverage because it can be monitored programmatically and cross-referenced against internal risk assessments as new incidents are added.
Context
AI Hack Watch collects public reports of AI involvement in hacking and security research. Its methodology separates the reported role of AI, such as autonomous action or human-directed assistance, from the severity of an incident. The timeline includes controlled research as well as misuse, so its entries do not all represent attacks on victims.
The site's counter measures calendar days since the newest selected article, generally using the source's publication date. It therefore tracks gaps between reports, not periods without attacks. Unreported incidents and events outside the site's selection are absent. The useful resource is the collection of linked evidence; the counter cannot establish how likely an organization is to be hacked.
Checking sign-in…
Loading comments…