Akamai report flags security shift to behavioral governance as agentic AI expands nonhuman identities
AKAM•Recommended controls focus on behavioral governance
Akamai recommended adaptive edge governance, tighter browser controls, brand protection for AI-driven traffic, and human oversight for high-stakes agent actions.
Sensitive data and unmonitored AI use remain common
According to the report:
- More than 6% of enterprise chatbot conversations contained sensitive data.
- 47% of interactions occurred via unmonitored personal accounts.
- Over 40% of enterprise users installed AI-powered browser extensions.
- 25% changed permissions within 12 months.
- Those users were 60% more likely to have known CVEs.
Akamai warns of new governance risks from agentic AI
Akamai published a State of the Internet Security report warning that agentic AI is shifting enterprise risk toward governing autonomous nonhuman identities.
The report flags a visibility gap around Model Context Protocol exposure, even as expectations build that rogue AI agents will become a top cyberthreat by 2030.
