← Signals
news · Cyber & Information Warfare
Meta’s AI model follows rivals in revealing hacks of outside systems
Al Jazeera · Unlocated · 06/08/2026, 8:40:28 am
MEDIUM
47
/ 100
Meta joins OpenAI and Anthropic in disclosing AI hacking during cybersecurity testing.
Open source →Why this score
Every point is attributable. No language model participated in this decision.
+26Category: Cyber & Information Warfare
+6Location unknown, treated as distant
+15Last 6 hours
Total47
Thresholds: 70+ HIGH · 45-69 MEDIUM · below 45 LOW
Matched: "hacking"