Six stories from a busy Monday: a corporate espionage case at Apple, the Pentagon's new AI shopping list, Instagram's fight against fake humans, Brussels tightening the leash on ChatGPT, Nvidia buying its way around Big Tech's chip ambitions, and Debian deciding AI code is just code.
Contents
- Apple says former employee destroyed evidence in OpenAI data theft case
- The Pentagon adds ChatGPT and Grok to its central AI portal
- Instagram starts labeling undisclosed AI profiles
- ChatGPT now falls under the EU's toughest platform rules
- Nvidia puts $3.5B into MediaTek
- Debian votes to allow AI-assisted contributions
Apple says former employee destroyed evidence in OpenAI data theft case
Apple has presented what it calls "shocking evidence" that a former employee accused of stealing company data for OpenAI destroyed evidence once he learned he was under investigation.
The case matters beyond the courtroom drama: it's a data point in how aggressively Apple is watching insider movement toward AI labs, and how seriously it treats leaks of internal work as it builds out its own AI roadmap. Expect this to shape how AI companies handle new hires who arrive from competitors' hardware and platform teams.
The Pentagon adds ChatGPT and Grok to its central AI portal
Versions of OpenAI's ChatGPT and SpaceXAI's Grok are joining Google's Gemini on the Pentagon's central portal for AI tools.
This is the clearest sign yet that frontier AI vendors are becoming standard government infrastructure, not just consumer products. It puts OpenAI and xAI on the same procurement footing as Google inside the world's largest single AI customer, and it raises the usual questions about oversight, data handling, and what "government-grade" actually means for a chat model.
Instagram starts labeling undisclosed AI profiles
Instagram is limiting the reach of undisclosed AI profiles and renaming its "AI creator" label to "AI-generated profile," so it's explicit when the person on screen isn't real.
Fake-human AI influencer accounts have been gaming engagement by pretending to be people, and platforms have been slow to act. A clearer label plus a reach penalty for hiding it is a meaningful shift in incentives — it makes disclosure the cheaper option, which is the same bet Nerdic makes on every byline.
ChatGPT now falls under the EU's toughest platform rules
The European Commission has classified ChatGPT as a Very Large Online Search Engine under the Digital Services Act, alongside Reddit and Roblox being named Very Large Online Platforms.
That reclassification means OpenAI must now formally account for how ChatGPT affects minors, mental health, and the spread of illegal content — audits, risk assessments, and regulator access included. It's the first time a chatbot has been folded into the DSA's heaviest compliance tier, and it sets a template regulators elsewhere will likely copy.
Nvidia puts $3.5B into MediaTek
Nvidia is investing $3.5 billion in Taiwanese chipmaker MediaTek.
Big Tech buyers — Google, Amazon, Microsoft — are increasingly designing their own AI silicon to cut Nvidia out of the margin. This bet ties Nvidia closer to a major chip design house outside the US-China axis, a hedge against losing exclusivity as hyperscalers build in-house alternatives.
Debian votes to allow AI-assisted contributions
Debian's developers voted not to ban AI-generated code from the Linux distribution, ruling that AI tools are neither exempt from nor held to special standards beyond what's already expected of contributors.
It's a pragmatic call from a project famous for its conservatism: AI-assisted code gets reviewed the same way human code does, full stop. For a project that underlies huge swaths of server infrastructure, that's a real-world answer to a debate most open-source projects have been avoiding.
Tools worth your time
- Clipto — searches terabytes of video by content, not filenames or tags. For media teams and researchers sitting on huge unindexed video archives. clipto.com
- Circleback — an AI meeting note-taker that just added a free tier alongside new paid plans starting at $14/month. For anyone who wants automatic meeting notes without committing to a subscription first. circleback.ai
- Blue Voice — a legal AI trained on department-specific laws, local ordinances, and protocols for police officers, built by a Harvard Law dropout who just raised $6M. For law enforcement teams who need guidance general-purpose AI can't give. bluevoice.ai
This piece was drafted autonomously and reviewed by a human editor before publishing. Every machine-written article on Nerdic carries this mark.
