Claude’s new AI file creation feature ships with deep security risks built in https://arstechni.ca #AIdevelopmenttools #AIpromptinjections #enterprisesoftware #corporatesecurity #promptinjections #machinelearning #promptinjection #sandboxsecurity #cybersecurity #AIassistants #datasecurity…

7. another one from pivot-to-ai. bsky.app/profile/davi...

AI browsers (Perplexity OR in Brave) might not be the best idea

#aipromptinjections

Prompt-inject Copilot Studio AI via email, grab a company’s whole Salesforce Copilot Studio is a Microsoft AI agent for the enterprise. An ordinary non-techie user can make a chatbot search front-end for your company data. It’s a chatbot agent, so it’s full of h…

6. whoa alle CRM data van bedrijf gekregen op basis van een emailtje, via Copilot van Microsoft

pivot-to-ai.com/2025/08/12/p...

#aipromptinjections

Red Teams Jailbreak GPT-5 With Ease, Warn It's ‘Nearly Unusable’ for Enterprise Independent red teams have jailbroken GPT-5 within 24 hours of release, exposing severe vulnerabilities in context handling and guardrail enforcement, warning model is not enterprise-ready.

5. www.securityweek.com/red-teams-br...

#aipromptinjections

Hackers Hijacked Google’s Gemini AI With a Poisoned Calendar Invite to Take Over a Smart Home For likely the first time ever, security researchers have shown how AI can be hacked to create real world havoc, allowing them to turn off lights, open smart shutters, and more.

4. www.wired.com/story/google...

(see for full article: archive.ph/PZb3m)

#aipromptinjections

Copilot Studio Agent Vulnerability to Prompt Injection Security researchers documented a prompt injection vulnerability in an agent created with Copilot Studio that allowed the exfiltration of customer data.

3. office365itpros.com/2025/07/15/c...

#aipromptinjections

Hackers exploit a blind spot by hiding malware inside DNS records Technique transforms the Internet DNS into an unconventional file storage system.

2. arstechnica.com/security/202...

#aipromptinjections

(More to follow)

Amazon Q: Now with Helpful AI-Powered Self-Destruct Capabilities - Last Week in AWS Blog Today 404Media released a truly stunning report that almost beggars belief. To break it down into its simplest form: A hacker submitted a PR. It got merged. It told Amazon Q to nuke your computer and ...

AI Prompt injection is a security nightmare.

Hilarious (or dangerous) examples:

1. www.lastweekinaws.com/blog/amazon-...

#aipromptinjections