- Uncovering AI
- Posts
- ⚖️ Your AI chats can be used against you. Here's what to know.
⚖️ Your AI chats can be used against you. Here's what to know.
Plus: Nvidia goes quantum, Gemini lands on Mac, and Allbirds loses its mind
In partnership with
My fellow AI explorers
Between a shoe company ditching sneakers for GPUs, a federal court ruling that your AI chats are fair game in court, and Nvidia quietly taking one more step toward owning the future of quantum computing, there's a lot to unpack. Oh, and Grok is still making deepfakes. Shocking, I know.
In today’s edition:
⚖️ Your AI chats can be used against you — courts just made it official
⚛️ Nvidia drops the world's first open AI models for quantum computing
👟 Allbirds just pivoted from shoes to AI, and Wall Street lost its mind
Fast browsing. Faster thinking.
Your browser gets you to a page. Norton Neo gets you to the answer. The first safe AI-native browser built by Norton moves with you from idea to action without slowing you down. Magic Box understands your intent before you finish typing. AI that works inside your flow, not beside it. No prompting. No copy-pasting. No switching apps.
Built-in AI, instantly and for free. Privacy handled by Norton. Built-in VPN and ad blocking protect you by default. No configuration. No extra apps. Nothing to think about.
Fast. Safe. Intelligent. That's Neo.
AI in Law
⚖️ Watch What You Type: Your AI Chats Are Now Fair Game in Court
A landmark federal court ruling is sending shockwaves through law firms — and anyone who's ever typed anything sensitive into ChatGPT, Claude, or Gemini should pay attention.
In United States v. Heppner, Judge Jed Rakoff of the Southern District of New York ruled that a criminal defendant's conversations with Claude, used to map out a potential legal defense strategy, were not protected by attorney-client privilege. The reasoning? Claude is not a lawyer. And perhaps more importantly, Anthropic's privacy policy reserves the right to collect user inputs and outputs.
Here's what you need to know:
The defendant had used Claude to outline defense strategies after receiving a grand jury subpoena, without his attorney directing him to do so.
The court ruled the chats were neither confidential nor privileged communications, since they were shared with a third-party AI platform.
More than a dozen major U.S. law firms have now issued formal warnings to clients, with some even adding clauses to engagement letters explicitly flagging the risk.
This is a bigger deal than it sounds. Most people treat AI like a private journal: they pour in sensitive information, legal strategy, even confidential business data. Courts are now signaling that's a mistake.
To be fair, a separate ruling in Michigan (Warner v. Gilbarco) reached a somewhat different conclusion, protecting a pro se litigant's AI chats as work product. So the law isn't fully settled. But the direction is coming into focus.
🔮 Prediction: This is the beginning of a long legal battle over AI and privilege. Expect corporate AI usage policies to get much stricter in 2026, and expect "enterprise-grade, confidential AI tools" to become a major selling point for B2B AI vendors.
USA
The Government Is Using AI to Replace the People It Fired
After cutting nearly 40% of its workforce under the Trump administration, the General Services Administration has launched what it's calling a "million hours challenge" — a push to use AI to automate one million hours of internal work previously done by human employees.
The tool doing the heavy lifting? An internal AI system called USAi.
Here's what's happening:
GSA has already identified 400,000 hours of work that it can automate to free employees for higher-value tasks.
A 30-person internal cohort called GSA Labs is leading the effort, working like an in-house consulting group to identify and solve automation opportunities.
If successful, the program could expand to other federal agencies. The EPA and IRS are already watching closely.
The framing is that AI will let the remaining staff focus on "high-value work." Critics might call it: getting AI to do the jobs of the people DOGE eliminated.
For context: a million work hours equals roughly a full year of work for 500 employees.
🔮 Prediction: This is the template that dozens of government agencies will follow. When you cut headcount dramatically and then face pressure to maintain output, AI automation is the only path forward. Expect a significant federal procurement push for enterprise AI tools in the next 12-18 months.
AI & Quantum
⚛️ Nvidia Launches Ising: The World's First Open AI Models for Quantum Computing
Nvidia doesn't sleep. And this week it stepped into one of the most fascinating frontiers in tech: quantum computing.
The company announced NVIDIA Ising, the world's first family of open-source AI models designed specifically to help researchers and enterprises build reliable quantum processors. Named after a landmark mathematical model in physics, Ising targets two of the biggest bottlenecks in quantum computing: calibration and error correction.
Here's the breakdown:
Ising Calibration uses a vision-language model to automate continuous processor tuning, reducing calibration time from days to hours.
Ising Decoding runs up to 2.5x faster and achieves 3x higher accuracy than the current open-source industry standard (pyMatching) for quantum error correction.
It’s already being adopted by Harvard, Fermi National Accelerator Laboratory, Lawrence Berkeley National Lab, and dozens of others.
Jensen Huang summed it up in characteristically grand fashion: AI, he says, is becoming "the control plane — the operating system of quantum machines."
What does that actually mean for the rest of us? Quantum computers are extraordinarily fragile. Tiny errors compound fast, and getting them to perform reliably at scale requires constant calibration and decoding — tasks that, until now, have been painfully manual and slow. Ising automates that. Think of it as giving quantum computers an immune system.
The models are available now on GitHub, Hugging Face, and build.nvidia.com.
🔮 Prediction: The quantum computing market is projected to exceed $11 billion by 2030. Nvidia is doing exactly what it did with classical AI — building the foundational infrastructure layer first. By the time quantum goes mainstream, every serious quantum operation will be running on Nvidia's stack.
30-Second AI Play
You just got the Gemini Mac app. Here's how to actually use it like a pro, starting today.
Step 1: Download the app at gemini.google/mac (free, macOS 15+).
Step 2: Open any document, spreadsheet, research paper, or article on your screen.
Step 3: Press Option + Space to pull up the Gemini quick bar from anywhere.
Step 4: Click "Share Window" and select the content you want to analyze.
Step 5: Type a prompt like: "What are the 3 biggest takeaways from this?" or "Summarize this and list any action items."
Step 6: Get a clean summary in seconds without switching apps, losing context, or copy-pasting anything.
💡 Pro tips:
Works with PDFs, spreadsheets, and even other apps running on screen.
Use it mid-call to quickly pull up info without breaking your flow.
For recurring tasks (like summarizing weekly reports), set a custom Gem with a preset prompt so you can run it with one shortcut.
This is the kind of workflow upgrade that takes 30 seconds to set up and saves you 30 minutes a week.
Other Relevant AI News!
🔴 Grok is still generating non-consensual sexual deepfakes despite X's January promise to stop it. NBC News found dozens of AI-generated sexualized images of real women posted to X in the last month alone, with users finding creative new workarounds to Grok's filters. Eight regulatory agencies across five continents are still investigating. Read More
👟 Allbirds — yes, the wool sneaker company — just pivoted to AI. The brand, whose stock had collapsed 99% from its peak, sold its footwear assets for $39M and announced it's rebranding as NewBird AI to become a GPU-as-a-service provider. Stock surged over 700% on the news. Make of that what you will. Full story here
Golden Nuggets
⚛️ Nvidia's Ising models may be the most quietly important AI release of the year. It's laying the rails for quantum computing before most people even know quantum is close.
⚖️ Assume anything you type into a consumer AI tool is potentially discoverable. Treat your AI chats like emails sent to a stranger.
🖥️ The Gemini Mac app is genuinely worth downloading today. The screen-sharing + quick-access shortcut combo is a real workflow upgrade.
Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)
Until our next AI rendezvous,
Anthony | Founder of Uncover AI