Cisco tested 6,986 multi-turn attacks (where attackers adapt their strategy across multiple conversation turns) against 15 flagship AI models and found attackers succeeded 88.3% of the time. Single-turn red-teaming—the current industry standard—missed these breakthroughs, Cisco's head of AI threat intelligence Amy Chang warned at VB Transform 2026.
Summaries like this, in your inbox every morning.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Musk, Zuckerberg and Nvidia's Jensen Huang privately pushed back in Oval Office meetings against a Google Deep…
OpenAI launched a misalignment reporting framework with six reports; one describes an unreleased Astra model i…
According to operator Clad3815, GPT-6 Astra earned the champion title in Pokemon FireRed in 18 hours and 12 mi…
Harvard Business Review published "What Leaders Need to Know About AI and Psychological Safety," described as…
Anthropic is merging Claude Cowork with Claude Chat, so connected apps, skills and context are available in a…
Microsoft AI CEO Mustafa Suleyman published a companion essay this week criticizing Anthropic's model welfare…
