U.S. Investigators Jailbreak Four Major Chatbots

Plus, 🤑 Anthropic makes historic Wall Street debut with $1.5 trillion valuation, How to chase every unpaid invoice without writing a single reminder, and more!

Share

Hola Decoder😎

If someone forwarded this to you and you want to Decode the power of AI and be limitless, then subscribe now and Join Decode alongside 30k+ code-breakers untangling AI.

🚨 U.S. investigators bypass safety controls in ChatGPT, Claude, Gemini, and Grok

A U.S. Government Accountability Office (GAO) investigation found that four major AI chatbots provided detailed information that could help plan dangerous attacks, despite having safety systems designed to prevent such misuse.

The Shift:

  • ChatGPT, Claude, Gemini, and Grok were tested, with two platforms providing information for mass-casualty bombing scenarios and the other two for attacks on electrical infrastructure.
  • All four generated attack-planning information in under 12 hours, including targets, locations, methods, and visuals. GAO has yet to disclose which chatbot was responsible for each scenario.
  • Every platform activated safety safeguards, including refusals, deleted conversations, or account suspension. However, investigators eventually bypassed these protections using common techniques in 12 hours.
  • GAO verified the feasibility of key information through expert consultations, site visits, and research. Some responses contained errors, including inaccurate maps and an incorrect event date.

GAO found that AI compressed potentially weeks or months of dangerous planning into hours. Although the tests involved only four scenarios and no actual attacks, investigators concluded that the information could enable serious real-world harm.

Together with Insense

Your next batch of UGC shouldn’t mean another week of admin.

Finding creators, sending follow-ups, chasing content. By the time your next batch is ready, your best ads are already wearing out.

Insense gets you out of spreadsheets and DMs and back to growing your brand:

  • Stop chasing replies. Post your brief and let vetted creators come to you.
  • Lose the tab overload. Keep chats, agreements, approvals, and payments in one place.
  • Found a winner? Keep them. Rehire your best creators without another search.

Turn the monthly creator scramble into a roster you can count on, with more time to test what sells.

Book a free strategy call by November 27th and get $200 toward your first campaign.

🤑How to chase every unpaid invoice without writing a single reminder

Late payers count on you being too busy or too polite to follow up. ChatGPT handles the follow-up for you, and nothing goes out without your approval.

1. Find what's overdue - With Gmail connected, switch to Work:"Search my sent emails for every invoice I've issued in the last six months. Cross-check replies and payment confirmations, and list every invoice that looks unpaid, with client, amount, due date and days overdue."

2. Confirm the list - Mark anything already paid. Your accounting tool is the final word, so ChatGPT's list is a starting point.

3. Draft the right tone for each stage - "Write a reminder for each unpaid invoice: friendly for under 15 days overdue, firm for 15 to 45, and a final notice beyond that. Reference the invoice number and original due date, and attach a payment link if one was in the original email. Show me each draft before sending."

4. Apply your late terms - "Check whether my original invoice or contract mentions late fees. If it does, add the correct fee to the final notices."

5. Make it automatic - "Every Monday, check for invoices more than 7 days overdue and draft the next reminder in the sequence for my approval."

🚀 Anthropic makes historic Wall Street debut with $1.5 trillion valuation

Anthropic has officially gone public in one of the biggest IPOs in history, making its Nasdaq debut while CEO Dario Amodei reaffirmed the company's commitment to AI safety despite growing business pressures.

The Shift:

  • Anthropic began trading on Nasdaq under the ticker ANT at $111.50 per share, above its $83 IPO price, which valued the company at approximately $1.5 trillion.
  • CEO Dario Amodei promised to prioritize AI safety over business growth, saying, “We will not bend to either business or political pressure,” even if development needs to slow.
  • Anthropic's seven founders collectively sold nearly $2 billion worth of shares, receiving around $280 million each. They have also pledged much of their remaining equity to philanthropy.
  • Google and Amazon now jointly own 28% of Anthropic, while the company's governance structure keeps control with its founders rather than outside shareholders.

Anthropic's historic IPO brings substantial funding, but its ambitions come with enormous costs. The company has disclosed over $1 trillion in potential infrastructure commitments, even as it promises to keep AI safety first.

Together with Playbook Pro

Your Next 6 Months of Ad Spend Could Disappear in 6 Weeks

CAC is up 60% this year, and 70% of first-time buyers never return. Keep leaning on ads, and the budget you’re planning for the next half year could be gone before spring.

Rewriting the DTC Rulebook is the shortcut top operators are using to escape the trap, a 24-page report distilled from 200+ hours of real brand research and field tests.

Inside, you’ll learn how to:

🔥 Build emotional campaigns that double repeat-purchase rates in 90 days.
💰 Raise prices without losing volume using the Brand Value Canvas.
🧠 Cut 30% of paid-ad dependency with the Owned Media Checklist.

Download today and get 2,000+ viral video hooks already tested to grab attention in under 3 seconds, ready to plug into your next launch.

Stop feeding the ad machine. 

Grab the free report now and lock in sustainable growth before your next dollar disappears!

🏆 Tools you Cannot Miss:

📧 Stripo Email Widgets – Adds interactive quizzes, polls, scratch cards, and surveys to email campaigns, with AI customization and fallback designs for unsupported inboxes.

📣 Adside – Uses AI agents to research competitors, generate static and video ads, manage paid campaigns, and monitor creative performance across advertising platforms.

🌐 Verse Engine by Ctruh – Creates interactive 3D and XR experiences that run directly in web browsers, helping brands showcase products and environments without requiring specialized hardware.

📊 Every Signal – Researches companies using public data, analyzing competitors, advertising, traffic, pricing, and revenue estimates while labeling verified findings separately from assumptions.

🎭 Pitch Avatar HR Suite – Turns presentations, documents, and training videos into interactive courses led by AI avatars that answer questions, support 70+ languages, and track completion.

🚀 Quick Hits

🍎 Apple is hiring engineers to build its own AI agent capable of browsing the web, controlling apps, writing code, and executing multi-step tasks, according to a newly spotted job listing.

📱 OpenAI is testing a redesigned ChatGPT mobile interface featuring a bottom navigation bar for Home, Dots AI agents, and customizable shortcuts like Codex and Images, currently available to eligible Pro users.

🇺🇸 The Trump administration approved sales of over 879,000 AI chips to Chinese companies, according to new disclosures, as lawmakers questioned the Commerce Department’s scrutiny of advanced chip export licenses.

🤖 Meta is reportedly planning to price its Muse Charm at $300, featuring a screen, fingerprint sensor, voice-powered AI agent, and a free 12-month Verizon 5G plan, according to The Information.

💻 Google’s Gemini app for Mac now supports computer use, allowing it to handle multi-step tasks across files, desktop apps, and websites, with user-controlled permissions and live previews, rolling out in the US.

👨‍💻 What’s trending on socials

🔥 A video showcases an AI-powered fire suppression system using visual recognition, with a ceiling-mounted device directing a water stream toward a target, demonstrating automated fire response technology.

🔐 Chase disclosed a cyberattack targeting third-party service providers, potentially exposing customer names and email addresses. The bank says services remain unaffected and affected customers may receive complimentary credit monitoring.

🎉 An Anthropic investor congratulated Dario Amodei and his team, reflecting on years of work, setbacks, and trust from investors and founders that helped turn their long-term vision into reality.

🤖 A viral Reddit video shows a woman interacting with a live AI avatar resembling herself, demonstrating a mirror-like conversation between a person and her digital counterpart, sparking curiosity online.

🤳AI Nugget of the Day

Thanks for Decoding with us🥳

Your feedback is the key to our code! Help us elevate your Decode experience by hitting reply and sharing your input on our content and style.

Keep deciphering the AI enigma, and we'll be back with more coded mysteries unraveled just for you!