AMD Watch posted an update
F-Secure and AMD Silo AI demonstrated a security architecture that decides, step by step, whether an AI agent's work runs on the device or in the cloud. It underpins TrustPath, an upcoming F-Secure platform protecting agent-driven browsing, logins and payments.
Why it mattersThe trick is judging each step of a journey rather than branding whole websites trusted or untrusted: a legitimate site can carry manipulated content or hidden instructions that fool an agent, and routine steps turn sensitive when credentials or payments appear. Routing weighs sensitivity, cost, model capability and performance, keeping processing close to the user where it can. TrustPath enters beta in Q4 2026, with production planned for 2027. These are vendor claims rather than independent tests, but the direction is clear: security is starting to decide where AI inference runs.
Discuss: If a legitimate website can carry hidden instructions that mislead an agent, does judging trust step by step contain that risk or just move it into the router?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.