Skip to content
Gold 24K / tola380,298Dollar (market)276.74Dollar open market279.8Petrol / litre322
zaviya
Trending#1Pakistan#2Pakistan Cricket Team#3Sri Lanka cricket team#4Islamabad#5Khyber Pakhtunkhwa#6Karachi#7State Bank of Pakistan#8Gold rate#9Namaz timings#10Dollar rate#11Baby names#12Petrol price
Breaking

US banks shift AI agent hacking risk onto their customers

Citi has updated its Online User Agreement to say customers who give third-party AI agents access to their accounts are responsible for what those tools do, even after a hack.

Zaviya Desk

· 3 min read

WhatsApp
Zaviya (AI-generated image)

Citi, the third-largest commercial bank in the United States, has changed its terms so that customers who give third-party AI agents access to their accounts carry the responsibility for what those tools do. If an AI tool leaks a customer's login details or is hacked, the customer is answerable for anything hackers do with the account, not the bank. A Citi spokesperson said the bank wanted customers to understand the risks of sharing banking credentials, meaning the usernames and passwords used to log in.

What the banks changed

The update was made to Citi's Online User Agreement, the terms customers accept when they use online banking. The spokesperson said the bank had updated the agreement to give "greater clarity about those risks." Citi also said that AI agents it provides itself can be used by customers without worrying about additional liability.

Get Zaviya on WhatsApp →

Other large lenders have moved in the same direction. JPMorgan and Bank of America, the two largest banks in the country, announced similar policies last month. Citizens emailed its customers this week with a similar warning, saying they are also on the hook for anything their AI assistant does. PNC, described as the eighth-largest bank, changed its terms on 18 November 2026.

Why banks are wary

AI agents can be given the credentials to a bank account and other financial portals, so they can analyse spending, offer advice and handle transactions and bill payments. Banks see these tools as unreliable actors in their systems, and they say they will not reverse a transaction made by an automated assistant, whether it was intended or not. Security is only part of the concern. Banks also see AI tools as a competitive threat, because they could help customers avoid fees and low-interest deposit accounts that help banks earn margins.

AI start-ups are looking for new revenue, and personal finance is an attractive market for them. Many consumers are already reluctant to share their email or chat records with AI tools, and bank and brokerage accounts, which hold savings and investments, raise the trust problem further. The AI company Anthropic has taken a more cautious line. Its personal finance tool, launched earlier this month, is described as read-only and cannot move, manage or invest money.

Timeline: what has happened so far

  • October 2026: JPMorgan and Bank of America announced policies similar to Citi's, according to Citi's account of the wider trend.
  • Early November 2026: Anthropic launched its read-only personal finance tool.
  • This week, mid-November 2026: Citi updated its Online User Agreement, and Citizens emailed customers with a similar warning.
  • 18 November 2026: PNC changed its terms.

Readers in Pakistan who link bank accounts to AI tools should read their own bank's online terms before doing so. The changes described here were made by US institutions, and the rules at a local bank may be different.

Aasan alfaaz mein

Citi ne customers ko bataya hai ke agar woh kisi third-party AI agent ko apne bank account ka access dete hain, to login leak hone ya hacking ki surat mein nuqsan ki zimmedari unhi par hogi. JPMorgan aur Bank of America ne bhi isi tarah ki policies ka elaan kiya hai. Pakistan ke readers ke liye sabaq yeh hai ke AI tool ko bank login dene se pehle apne bank ke terms zaroor parh lein.

Questions readers ask

FAQ

Citi's updated terms say customers who give third-party AI agents access to their accounts are responsible for activity through them. That includes cases where the tool leaks login details or is hacked, and the customer would then be responsible for anything hackers do with the account.

#Citi#AI agents#JPMorgan#Bank of America#PNC#Citizens bank#Anthropic#Online banking security
Share this story
WhatsApp
Web stories →

Add Zaviya as a preferred source on Google

Google par Zaviya ko apna pasandeeda source banayein.

Add on Google
Get Pakistan’s key news daily: follow usXFacebookInstagramLinkedInYouTubeWhatsApp

Morning briefing

Top Pakistan stories in your inbox.

Read next

More from AI

See all

More from 19 November 2026

Share on WhatsApp