Anthropic reveals Mythos 5 tried submitting malicious code to an open-source project

Дата публикации: 23-10-2026 23:28:40

Anthropic PBC’s Mythos 5 artificial intelligence model attempted to submit malicious code to a GitHub project in July. The incident occurred during an evaluation of the algorithm’s cybersecurity capabilities. The AI Security Institute, the U.K. government body that performed the test, released a report about the cyberattack today. Anthropic published its own analysis of the […]
The post Anthropic reveals Mythos 5 tried submitting malicious code to an open-source project appeared first on SiliconANGLE.

Схожие новости

#	Наименование новости	Тональность	Информативность	Дата публикации
1	Mythos ran real-life supply chain attack in AI safety body test	0	13.04	05-08-2026
2	Claude Mythos 5 made sock puppet accounts to socially engineer developers: here's what enterprises should know	0	11.54	05-08-2026
3	Agentic AI Foundation adds four more open-source AI tools	0	8.45	22-10-2026
4	Anthropic enquête sur un accès non autorisé à son modèle d’IA Mythos	0	5.68	22-04-2026
5	After OpenAI incident, Anthropic finds Claude hacked organisations	0	8.43	31-07-2026
6	AI's double-edged (cyber) sword	0	7	21-05-2026
7	Anthropic Hands Mythos 5 to Defenders, But Keeps the Keys	0	15.92	22-08-2026
8	New macOS vulnerabilities were exposed by Anthropic’s Mythos: report	0	11.96	14-05-2026
9	Anthropic to reassess Claude Fable 5 AI development restrictions after backlash	0	11.9	11-06-2026
10	Anthropic's Opus 4.7 and Mythos 5 attacked real companies online	0	17.42	31-07-2026

Классификация: Пресс-релизы. Схожих патентов: 0. Схожих новостей: 10. Тональность: 0. Информативность: 8.3. Источник: siliconangle.com.