Security | VentureBeat

AI agents are about to put least privilege to the test

Tue, 20 Oct 2026 14:30:00 GMT

Presented by JumpCloud

DeepSeek-V4.1-Flash-1017 arrives with improved benchmarks just as the U.S. government bans it from agencies — what enterprises should do

carl.franzen@venturebeat.com (Carl Franzen) — Mon, 19 Oct 2026 23:46:38 GMT

DeepSeek keeps giving businesses reasons to try its low-cost, high-performing, open source AI models. The U.S. government, though, is saying 'no dice.'

AI cloud security needs to prove control before agents reach production

Mon, 19 Oct 2026 20:52:50 GMT

As confidential computing comes to GPUs, it can protect sensitive workloads, but enterprises also need evidence that they can identify agents, isolate tenants and revoke access without disrupting operations.

Anyone can build apps with AI. But who owns them?

Mon, 19 Oct 2026 15:00:00 GMT

Presented by Retool

17 of 40 teams that say they enforce AI agent permissions still share the credentials behind them

louiswcolumbus@gmail.com (Louis Columbus) — Fri, 16 Oct 2026 13:00:00 GMT

VentureBeat’s September 2026 Agentic Security and Identity survey asked 102 enterprises how they secure and revoke AI agent access. Forty described a program that enforces scoped identities and runtime permission checks, and 17 of those 40 then said some or most of their agents still share credentials. Only 14 of the overall 102 said they could revoke access within minutes.

Browser security’s buying spree rests on one engine, without a clear Plan B

Wed, 14 Oct 2026 14:30:00 GMT

Presented by CloudMosa

Anthropic’s Cyber Mission starts with 6,157 findings reported to maintainers and 516 patched

louiswcolumbus@gmail.com (Louis Columbus) — Fri, 09 Oct 2026 07:00:00 GMT

Anthropic’s models have outrun the people who patch what the models find.