Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
AI Businesses

Bankers Will See AI Transform Three-Quarters of Day, Study Says (bnnbloomberg.ca) 36

AI is likely to replace or at least lend a hand in tasks that take up almost three-quarters of the time bank employees now spend working. From a report: That's the conclusion of a new analysis by consultancy Accenture, which said banking has the potential to benefit more from the technology than any other industry. Just 27% of employees' time currently has a low potential of being transformed, according to the analysis. "There is a reinvention that is happening across banks, a way for firms to step back and re-evaluate ways of working," Keri Smith, global banking data and AI lead at Accenture, said in an interview.

The release of ChatGPT more than a year ago prompted many firms to boost hiring for AI-related positions and test more uses for generative AI, which can summarize documents, write emails and churn out responses to users' questions. The world's biggest banks have been experimenting, spurred by the promise that the technology will boost staffers' productivity and cut costs. "Every bank needs to think through their talent strategy, and how to take this technology to scale," Smith said. At Citigroup, all 40,000 coders will have the ability to experiment with different AI technologies by the end of March. Analysts at Bank of New York Mellon can wake up two hours later to write their research, because AI technology can create a rough draft and prepare related data for them overnight, Chief Executive Officer Robin Vince said on an earnings call last month.

This discussion has been archived. No new comments can be posted.

Bankers Will See AI Transform Three-Quarters of Day, Study Says

Comments Filter:
  • 3/4 (Score:5, Insightful)

    by groobly ( 6155920 ) on Tuesday February 27, 2024 @11:44AM (#64272784)

    From now on, bankers will spend 3/4 of the day fixing problems caused by reliance on AI.

    • Re:3/4 (Score:5, Insightful)

      by fleeped ( 1945926 ) on Tuesday February 27, 2024 @11:48AM (#64272788)
      Step 1. Chatbots says something ... importantly wrong
      Step 2. Bank gets sued
      Step 3. Bank limits AI to "safe levels", that cripple it and make it effectively useless (a hallucination machine that can't hallucinate)
      Step 4. BINGO! Back to where we are, with useless chatbots.
      • by Tablizer ( 95088 )

        Chatbots says something ... importantly wrong ... Bank gets sued

        "No, Mom is NOT classified under 'cattle' on my taxes, Idiot BankBot!"

        • No, damn robot, my golf course cannot be classified as a cemetery in this years tax form! Stupid chatgpt... Oh, wait!

    • Given the banking industry's history of racial bias, I sure hope somebody's looking out for the 3/4ths folks. Depending on what they're ingesting for source material into the large language model, redlining could come back by accident fairly easily.

  • Everyone knows you need at least 6/8 to have it make any sense.
  • ...this is great, with competent AI robots accurately doing tasks that people make mistakes doing
    In practice, it will be another instance of enshittification, making customers even more angry as stupid robots make their lives even more miserable
    I still believe that future AI will be good on balance, but the transition will be painful

    • Re:In theory... (Score:5, Insightful)

      by serviscope_minor ( 664417 ) on Tuesday February 27, 2024 @01:24PM (#64273036) Journal

      ...this is great, with competent AI robots accurately doing tasks that people make mistakes doing

      Even assuming your optimistic scenario (not the in practice one) pans out bear in mind this:

      The release of ChatGPT more than a year ago prompted many firms to boost hiring for AI-related positions and test more uses for generative AI, which can summarize documents, write emails and churn out responses to users' questions.

      If the document can be reasonably summarized, why was such a long one written?

      And if the AI can write emails, presumably from a short prompt, then why does it need to be longer than that prompt in the first place?

      The implication that they haven't admitted to is that there was an awful lot of shit and busywork being done making the recipients lives harder by having to wade through the busywork. Rather than doing the obvious thing of removing the shit, they are making tools to generate the shit more efficiently.

      Even your optimistic scenario is bad.

      • You know that some lawyers will worry that the summarized document isn't complete enough and will insist it is a little longer. After a few iterations, the summary document will need a summary document. The customer will eventually just be reading a summary of a summary of a summary of a summary of a summary of a summary of a summary of a summary of a summary. Why can't the customer understand their legal obligations?
  • I could see the banking industry creating their own AI

    • With management like this... get out now!
      Accenture changed their name they screwed up their job so badly or corruptly!

      So they can do more massively huge fraud and now blame the AI; although, nobody really gets in trouble, so why bother? oh yes, it's raising stock prices like mentioning blockchain did...

    • They absolutely would trust someone else's AI. Banking is very willing to use software products made by all kinds of vendors. I saw this first-hand from working in mortgage processing. A single mortgage may be processed by software from 50-100 different software vendors, including those that claim "AI" capabilities.

  • by cuponadesk ( 6805092 ) on Tuesday February 27, 2024 @12:21PM (#64272868)
    Iâ(TM)m skeptical that AI will play any major part at banks anytime soon. Banks have very large regulatory and compliance mechanisms that are ingrained in every customer facing interaction. There is zero percent chance AI is allowed to interact with anyone outside of what is already possible with standard script-based automation. The one area I could see it being helpful is documenting existing code. A lot of banks struggle with legacy code that is so poorly documented that no one is exactly sure how its logic works. The other issue is data retention. Improperly configured LLMs could potentially retain and regurgitate personal identifying information, which would lead to a regulatory and compliance nightmare. Even then are you going to section off individual applications from each other so that only authorized users can get info they have access to read? I highly doubt many places have an interest in setting up a segregated AI for every squad or group.
  • Very inefficient mode of theft.
  • What can these excess bankers do that AI cannot? Direct traffic and fill pot holes? The possibilities are endless.

  • Loan application interview by ai... Ai: Any other information pertinent to your application? Applicant: yes. Given that i am wealthy and own substantial land there is no risk in approving my application so it should be approved immediately. Ai: Application approved!
  • Compute power gave us the derivatives that caused the 2008 crash by splitting up debts into untraceable atoms. What will AI bring to finance but more ways to obfuscate fraud and risk? The last thing society should want is for AI to play any sort of significant role in finance. This is not going to end well. Needs regulation ASAP.
  • I hope, for bankers' sake, that the three-quarters of each day that AI revolutionizes overlaps with the three-quarters of each day the bank is already closed.
  • Bankers Will See AI Transform Three-Quarters of Positions To Redundant, Study Says

  • "Global markets are in freefall today as chat bots hallucinate amount of money in bank accounts."

  • "AI is likely to replace or at least lend a hand in tasks that take up almost three-quarters of the time bank employees now spend working"

    =

    Banks announce 75% employee layoffs.

    Some of us are old enough to remember when you went into a bank, ALL THE TELLER WINDOWS HAD HUMANS. ....until ATMs.

  • Oh great, now visits to the bank will be like this [youtube.com].

If all else fails, lower your standards.

Working...