While other financial institutions are banning ChatGPT, Goldman Sachs revealed that it is changing course and will use generative artificial intelligence (AI) tools to write and test software code.
Key Details
- Generative AI has been in use as a “proof of concept” and is not ready to be rolled out, Goldman’s chief information officer Marco Argenti told CNBC.
- After previously not permitting its use, now around 40% of Goldman’s code has been written by the AI tool as software developers use it to expedite their work.
- Argenti emphasized that Goldman does not intend to replace human workers with AI but rather use it to enhance a worker’s ability.
- Argenti did not reveal in his comments whether or not Goldman is using AI in its investment bank.
Why it’s news
Goldman’s willingness to reverse course and integrate AI technology into its systems stands out as to how rapidly companies are embracing the technology. JPMorgan notably restricted employee access to ChatGPT last month. While a particular incident did not cause the ban, JPMorgan officials made this decision out of apparent concern about the potential for data leaks.
Other large banks, such as Wells Fargo, Citigroup, Bank of America, Deutsche Bank, and Goldman, have also restricted employee access to ChatGPT, Forbes reports. These restrictions were primarily part of standard procedures that limit employee access to third-party software.
A growing interest in generative AI has companies from all industries exploring how the new tool can be applied in their respective situation. Goldman’s method of using AI to write code is one of many suggested uses for AI.
Argenti did not reveal which generative AI products Goldman has been using or which bank division the tool is being used in.
Backing up a bit
JPMorgan Chase is limiting its staff’s access to ChatGPT. Amazon and multiple U.S. universities have also restricted the use of the AI chatbot.
The financial firm’s decision to block the AI wasn’t caused by any particular event, but it appears to be a preemptive move to prevent potential security leaks. The company says the limits are part of JPMorgan’s “normal controls around third-party software.” JPMorgan’s decision may have also been motivated by concerns that sensitive information would be shared with ChatGPT, the Telegraph reports.
Amazon has also warned its employees about including sensitive information in any conversations with the chatbot.