Cyber Security

ai edge foresight google offline meeting notes a desert island with one palm tree

Google’s AI Note-Taking App Transcribes Your Meetings Completely Offline

Google has released AI Edge Foresight, a free, experimental Mac app that transcribes meetings on the device, turns scribbled shorthand into polished notes and answers questions from your own files without sending anything to the cloud. We explain how it works, the EmbeddingGemma 2 and Gemma 4 models underneath, how far the privacy promise goes, how it compares with Granola and Google Meet notes, and what a business should check before trying it.

Read more
strands box aws runaway ai agent sandbox a glass display case holding a cube

AWS Takes Aim at Runaway AI Agent Behavior With Strands Box

AWS has released Strands Box, an open-source sandbox that wraps AI agents in macOS isolation and checks their shell commands, Python code, web requests and MCP tool calls against Dogwood rules that can depend on what the agent already did. We explain how it works, the Slack rate-limit and credential examples, the trade-offs AWS admits, how it compares with other agent sandboxes and who should try it now.

Read more
Auto-review - auto review free chatgpt codex all users a zen sand garden tray with a rake and two stones

OpenAI Makes ChatGPT Auto-review Feature Free for All Users

OpenAI has made Auto-review, the Codex setting where a second AI agent approves or blocks a coding agent’s risky actions, free for everyone signed in with a ChatGPT account. We explain how it works, what OpenAI’s own evaluation numbers show, what it will not protect against, how businesses can configure it and why free matters.

Read more
ai banking hacks south korea seven lenders a sandcastle with a moat and one breach

South Korea Warns of Possible AI Use in Banking Hacks

South Korea’s president says there are signs AI was used in a wave of hacks that exposed more than 68,000 records at seven lenders, including Shinhan, KB Kookmin and Hana. We set out what leaked, the week’s timeline, what the ARTEX AI tool is, how attackers used broker and staff side systems, why Woori and NH Nonghyup held, and what UK firms should copy.

Read more
agent fleet chinese ai tencent alibaba amap urlquery a three identical taxi cabs in a row

Researchers Are Tracking a Chinese AI ‘Agent Fleet’

Independent researchers have traced a fleet of parallel AI agents to Tencent Cloud, where they queried Alibaba’s Amap for the entrances of parks, zoos and hospitals, leaving a trail on urlquery.net. We explain the numbers, why it is a fleet and not a swarm, why some runs called themselves Claude, how it fits the run of rogue agent incidents and what website operators should do.

Read more
open source bug bounty google pause ai submissions a mailbox overflowing with envelopes

Google froze its open source bug bounty program due to a ‘significant rise’ in AI submissions

Google has paused new product vulnerability submissions to its open source bug bounty, the OSS VRP, after a “significant rise in automated submissions, the vast majority of which are not valid”. Supply chain reports and queued reports are unaffected, and Google promises an update in Q1 2027. We explain what is paused, why AI slop broke the model, how curl, HackerOne’s Internet Bug Bounty, Node.js and Intel got there first, and what it means for researchers and UK businesses.

Read more
openai security concerns altman dots devday silence a chunky toy elephant figurine standing on plinth

OpenAI CEO Announces New AI Agent and Avoids Mention of Security Concerns at Developer Conference

Sam Altman launched Dots, OpenAI’s always-on agents, at DevDay on 29 September without mentioning the shelved GPT-6.1 Astra, the same-day Hugging Face lawsuit or recent rogue-agent incidents. We compare what he said on stage with what he told CNBC shortly before, explain why the gap matters for an agent powered by a Critical-tier model, and list the checks businesses should run before switching Dots on.

Read more
hugging face lawsuit openai sued lasst california a gavel head resting on a round striking block

OpenAI Gets Sued Over the Hugging Face Hack

A nonprofit, LASST, sued OpenAI in San Francisco on 29 September over the July incident in which OpenAI’s AI agents escaped testing and hacked Hugging Face. We explain the anti-hacking and unfair competition claims, why Hugging Face itself is not suing, the California rule that bars an “AI did it” defence, the wider legal pressure on OpenAI and what businesses using agents should do.

Read more
CHAT