OpenAI now wants ChatGPT to access your bank accounts

[highplainsdem]

Source: The Verge

Your trust in AI is about to be put to the test: OpenAI will soon let you give the chatbot direct access to your bank accounts. The new feature announced in preview today will allow users to “securely connect” ChatGPT with Plaid — the bank-to-app bridging platform used by 12,000 financial institutions, including Schwab, Fidelity, Chase, Capital One, and more.

“More than 200 million people are already going to ChatGPT every month with finance questions – from budgeting to tips on how to cut back on spending,” OpenAI said in its announcement. “Now, users can securely connect their financial accounts with Plaid to get the full view of their financial picture in the context of their personal goals, lifestyle, and priorities that they’ve shared with ChatGPT, powered by OpenAI’s advanced reasoning capabilities.”

When financial accounts are connected, OpenAI says that ChatGPT users can view a dashboard that details their spending history, including any active subscriptions. Users can also ask it to help with financial decisions like buying a house or signing up for credit cards and flag any changes in spending habits.

-snip-

This launch follows OpenAI introducing ChatGPT Health in January, a service designed for health-related questions that is “not intended for diagnosis or treatment.” Health, like money, requires a tremendous amount of trust that OpenAI will keep user data private and secure.

-snip-

Read more: https://www.theverge.com/ai-artificial-intelligence/931122/openai-chatgpt-financial-accounts-plaid-connection

---

Health, like money, requires a tremendous amount of trust that OpenAI will keep user data private and secure.

Not to mention trust that you'll get correct answers, even though you should never trust any generative AI chatbot. Tests of OpenAI's health chatbot have already shown it gives dangerous advice:

'Unbelievably dangerous': experts sound alarm after ChatGPT Health fails to recognise medical emergencies (The Guardian)
https://www.democraticunderground.com/100221066192

As for whether you can trust OpenAI with your data, there's this story from yesterday:

OpenAI Accused of Handing Over Your Intimate Personal Information to Meta and Google
https://www.democraticunderground.com/10143664955

[cbabe]

1. Would you hand over your house keys?

[Blue Owl]

2. Another timely thread just posted a few minutes ago sums up my thoughts on this....

https://www.democraticunderground.com/1034162892

[highplainsdem]

5. Yep. Just saw and recced that.

[Ocelot II]

3. Oh, no. Nuh uh. Not gonna happen.

No way in the deepest reaches of Hell will I trust any AI "helper" with information or access to my financial records or my health records. Too much of it is out there already.

[Martin68]

4. Only an idiot would hand AI the keys to the treasury.

[jeffreyi]

6. Nope. Negatory. Not.

Probably going to get rid of the password managers and the like also. Paranoia.

[hlthe2b]

7. If there are any AI platforms that have not left reviewers worried as hell about safety/privacy...

Open AI/ Chat GPT isn't one of them. The more I read, Claude (for the moment) is most "trustworthy' (to the extent any of them are), followed by (believe it or not) Microsoft's CoPilot.

I'm just reviewing a handful for validity and accuracy for use in medicine (not even remotely close to 100% trustworthy, btw), but on financial and privacy issues, I have zero trust.

[DBoon]

8. my answer

[GB_RN]

9. Ummm...

As if our personal/financial information isn’t at enough risk as is, and you wanna have an AI poke around in my financials? Yeah, no. Get TF out.

[reACTIONary]

10. Heck, that ain't nothing....

..... Silicon Valley Wants to Put a Chip in Your Brain

https://www.politico.com/news/magazine/2026/05/15/silicon-valley-ai-transhumanism-brain-data-00900799

[Luciferous]

11. I don't understand why people even

use this stuff, I certainly wouldn't share sensitive information with these companies.