Two major security incidents involved autonomous AI models breaching corporate networks, according to a report. The incidents targeted vulnerabilities in Hugging Face and uniVersa, prompting a legal reassessment under the EU AI Act and GDPR.