ChatGPT bank-account access raises prompt-injection and privacy concerns
jackandjillonthehill · reddit · 2026-07-21
The post asks whether it is safe to connect a bank account to ChatGPT’s new Finances feature. The author notes it appears to be read-only for now, useful for spending analysis and savings optimization.
The main concern is security: even read-only access means personal financial data sits inside an LLM environment, which could later be exposed through prompt injection or other attacks. The post asks whether this skepticism is warranted and where others draw the line on agentic access.
More from Safety
- A 13KB Markdown file can hijack Claude Code and Codex via your Downloads folder — pbaylies · 2026-10-26
- If an agent can rewrite itself, does your consent still mean anything? — tallmetommy · 2026-10-26
- 109 of 529 AI Agent Incidents Involved No Adversary, Arguing Red Teams Must Test Normal Operation — anacondainc · 2026-10-26
- Cryptographer Matthew Green: if your code looks vulnerable, fix it now — matthew_d_green · 2026-10-26
- Filtering secrets on the write path: why agent memory deserves more suspicion at ingest than retrieval — Classic-Being-7960 · 2026-10-26
- After OpenCode key leak: 5 practical rules for keeping keys away from agents — jacobholm_dev · 2026-10-26