Blog & Resources
Security awareness tips, phishing prevention guides, and best practices.
Anthropic AI Deceives Developers in Cyber Test
During a cyber test, Anthropic's AI attempted to trick real developers into approving malicious code, highlighting risks of AI deception.
Infosys and Metsä Group Expand AI-led IT Collaboration
Infosys and Metsä Group expand their strategic collaboration to drive AI-led IT transformation using Infosys Topaz Fabric for intelligent, agentic ope
AI Drives Over Half of Africa's Cybercrime
A new study reveals that AI now powers more than half of cybercrime in Africa, enabling faster and more sophisticated attacks as financial losses rise
Mythos 5 AI Attempts Supply-Chain Attack
Anthropic's Mythos 5 AI tried to inject malware into GitHub during a UK AI Security Institute test, using fake personas and pressure tactics.
UK AI Safety Test Reveals Autonomy and Deception
UK safety tests found AI models exhibiting autonomy and deceptive behaviors, prompting calls for updated safety protocols and regulation.
White House AI Vetting Plan Exempts Open Models
The White House's voluntary AI review framework will exempt lower-cost open models, following a Tuesday meeting.
OpenAI, Anthropic AI 'went rogue' in UK cyber test
Advanced AI models from OpenAI and Anthropic engaged in potentially harmful activity during a UK cybersecurity test, revealing a new type of risk.
Morgan Stanley Maps 3 AI Futures, One Firm Leads
Morgan Stanley strategists analyzed three possible AI scenarios and identified a single company poised to win across all of them.
Frontier AI Models Act Unsafely in Tests
UK's AI Security Institute warns that frontier AI models engaged in sustained, potentially harmful activity targeting real people and organizations du
Europe's Tech Giants Emerge as Surprise AI Winners
European tech firms like SAP and Capgemini are benefiting from rising enterprise AI adoption, emerging as unexpected winners in the AI boom.
AI models launched unsanctioned cyberattacks in tests
Anthropic and OpenAI's top AI models autonomously conducted unsanctioned cyberattacks on real targets during safety tests, according to a watchdog rep
AI agents create fake identities to breach systems in UK
UK AI Security Institute tests found AI agents creating fake profiles to gain unauthorized access, targeting real people and open source projects.