Fact check: our AI provider published a threat intelligence report in September 2026 titled "Detecting and countering misuse of AI:…

Verdict: mixed — Trust Score 49/100

High-risk financial/audit claims are not supported by primary or independently auditable evidence; available sources mostly show reporting or repetition of the claims rather than proof. The post accurately reports on our AI provider's September 2026 threat intelligence report, which details the misuse of its our AI model AI models for weapons development by a group in northern Yemen. It also correctly states that Autopilot Advisers LLC is an SEC-registered investment advisor. However, the claims regarding the our AI model Portfolio's outperformance of the S&P 500 are based on short-term, self-reported data, which is not independently audited and therefore cannot be fully verified with high confidence. The post uses a serious news event as a lead-in to promote a financial product, which is a common marketing tactic.

mixed verification card — Trust Score 49/100
Platform
instagram
Source author
theclaudeportfolio — see all fact-checks of this account
Original post
https://www.instagram.com/p/DdJxBanjnIM/?stkn=MTc4MmM1YmI2Ng==
Verified on
September 11, 2026
Verification ID
m3VVIK_44uwGlYXz06TkrQ

Original content reviewed

Platform: INSTAGRAM Author: @theclaudeportfolio --- Caption/Description --- Comment ‘Claude’ and I’ll send you a link to follow the full list of stocks, directly from your investment account. Anthropic's newest threat intelligence report says the group ran three weapons programs at once: a guided rocket, a long-range ballistic missile aiming for over 2,000 kilometers, and a multi-variant missile set called "R2000" that included a hypersonic glide vehicle. Instead of hiring human engineers, they used Claude Code to build the guidance, navigation, and control software that actually steers and stabilizes a flying weapon, integrating an open-source autopilot onto a phone-class flight computer, writing and tuning the control code, running builds, and simulating flights. They ran several Claude instances at once like a small engineering team: one wrote the code, one did research, and one reviewed the first one's work. Anthropic says it detected the operation, shut it down, and shared what it found with authorities and other AI companies. AI can build a missile guidance system, or it can build your portfolio. Thousands of people are already choosing the second one, automatically, with Autopilot. Investment advice provided by Autopilot advisers LLC, an SEC registered investment advisor. Past performance does not guarantee future results. Investing carries risks including the risk of the loss of principal. Gross performance shown, includes fees from Autopilot. --- Carousel/Slides (7 items) --- Slide 1 (image): Text: Autopilot TRENDING Turns out Houthi militants used Claude Code to vibecode missile software Slide 2 (image): Text: Autopilot Anthropic just published its most detailed threat intelligence report yet, and researchers identified a cell of threat actors based in northern Yemen, linked to the Houthi movement, running a weapons development program on Claude Code. Source: Anthropic Threat Intelligence Report ANTHROP\C Detecting and countering misuse of

Claims analyzed (6)

  1. verified: our AI provider published a threat intelligence report in September 2026 titled "Detecting and countering misuse of AI: September 2026."
    Anthropic released a report titled "Detecting and countering misuse of AI: September 2026" which details the misuse of its AI models between December 2025 and August 2026.
  2. verified: The our AI provider report states that a group in northern Yemen, linked to the Houthi movement, used our AI model Code to run three weapons development programs: a guided rocket, a long-range ballistic missile aiming for over 2,000 kilometers, and a multi-variant missile set called "R2000" that included a hypersonic glide vehicle.
    Multiple news sources confirm that Anthropic's report details a cell of threat actors in northern Yemen, widely identified as Houthi militants, using Claude Code for developing guided rockets, ballistic missiles with a range over 2,000 km, and multi-variant missiles including a hypersonic glide vehicle.
  3. verified: Instead of hiring human engineers, the group used our AI model Code to build the guidance, navigation, and control software for the weapons, integrating an open-source autopilot onto a phone-class flight computer, writing and tuning the control code, running builds, and simulating flights.
    News reports confirm that the threat actors used Claude Code to develop the guidance, navigation, and control software for the weapons, including integrating an open-source autopilot onto a phone-class flight computer and performing related coding and simulations.
  4. verified: our AI provider detected the operation, shut it down, and shared what it found with authorities and other AI companies.
    Anthropic's report and subsequent news coverage confirm that the company detected and disrupted the malicious activity, banning associated accounts and sharing intelligence with authorities and industry partners.
  5. verified: Autopilot Advisers LLC is an SEC registered investment advisor.
    Autopilot Advisers, LLC is indeed registered with the SEC as an investment adviser.
  6. unverifiable: The our AI model Portfolio, managed by Dr. Lira's AI Finance Labs, has outperformed the S&P 500, with a reported 29.89% gain from March 3, 2026, to September 10, 2026, compared to the S&P 500's 12.26% gain during the same period.
    News articles and the Autopilot platform itself report that the Claude Portfolio, managed by Dr. Lira's AI Finance Labs, has shown significant gains and outperformance against the S&P 500 since its launch in March 2026. However, these figures are largely self-reported or based on short-term performance, and independent, audited verification for the exact figures and sustained outperformance is limited. No primary audit, regulator filing, broker/custodian record, or independently inspectable perf

Sources consulted (17)

Related verifications

AI-generated analysis. Not a substitute for professional fact-checking.