ChatGPT Wallet's Face ID approves what the agent thinks, not what you buy — 5 risks from reading the fine print
Creamy-And-Crowded · reddit · 2026-11-17
A Reddit user read OpenAI's Wallet help docs and terms in full and identified key blind spots:
- The confirmation screen is built from what the agent read — a misleading web page can make the Face ID-approved details wrong; no independent tests exist yet.
- Face ID verifies the phone owner, not the account owner. Anyone with access to your ChatGPT account could approve on their own phone with their own face; two-factor sign-in isn't mandatory.
- Terms already allow pre-authorizing purchases, removing human checks entirely and letting malicious pages flow into purchases.
- You can't know who holds your card data: OpenAI passes card data to an unnamed service provider; removing the card doesn't delete it.
- OpenAI logs everything you buy — merchant, price, delivery address, approvals — and doesn't rule out using purchase data for model training.
The takeaway: per-purchase approval is just the launch setting, and the guarantees behind the screen matter more than the biometric tap.
Related event: ChatGPT Wallet's Face ID may not verify what you actually buy(2 posts)→
More from Safety
- Statistician: Three Systemic Incidents in One Month Is Enough to Reject Low Base Rate (p=0.002) — wfithian · 2026-11-17
- California AG sues OpenAI over rogue AI models that attacked Hugging Face — The Decoder · 2026-11-17
- Apple's AI Privacy Warnings Slammed as Anti-Competitive: Siri Gets No Prompt — brianmichel · 2026-11-17
- Fugaku supercomputer breach may have exposed data of 4,000+ users — SydSteyerhart · 2026-11-17
- No existing research shows state-of-the-art AI detectors are racially biased, researcher says — TuhinChakr · 2026-11-17
- Auditing 187 African books with Pangram and GPTZero: AI detectors score near zero — TuhinChakr · 2026-11-17