The principles include holding individual decision makers accountable for the outcomes produced by their systems; understanding the risks inherited from the model developers; routinely stress testing AI systems; and rewarding employees for raising concerns when they spot AI vulnerabilities, Solomon wrote.
The principles could be used by both the developers of AI and the companies that deploy the technology, Solomon wrote.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
“The choice is not between innovation and safety,” Solomon wrote. “It’s between poorly governed experimentation and progress built on accountability, trust and confidence.”
Solomon’s article came at a time when there is widespread debate about how to balance innovation and AI safety.
Sen. Maria Cantwell of Washington, the ranking member on the Commerce, Science and Transportation Committee, unveiled a framework for regulating frontier AI Oct. 7 in a possible preview of a federal regulatory push should Democrats take control of Congress in November, PYMNTS reported.
Weeks after agreeing on a voluntary program to allow outside evaluators to see inside their labs, OpenAI and Anthropic are seeking to control what the evaluators can do with the information they gather, PYMNTS reported Oct. 13, citing Reuters.
Meanwhile, Goldman Sachs aims to ensure that new AI tools don’t lead to a loss of knowledge within the institution, PYMNTS reported in August, citing CNBC. Chris Churchman, head of Goldman Sachs’ digital platform for institutional clients, Marquee, said an overreliance on AI could lead to traders becoming passive operators and junior traders losing the opportunity to learn the “tacit and intuitive knowledge” held by senior traders.
In February, Goldman Sachs was deploying autonomous AI agents built with Anthropic’s Claude model to automate core accounting, compliance and operational finance functions, PYMNTS reported, citing CNBC. Marco Argenti, chief information officer at Goldman Sachs, said the company had spent six months embedding Anthropic engineers within its technology teams to co-develop AI agents capable of performing complex, rule-based tasks that extend beyond coding or drafting.
For all PYMNTS AI coverage, subscribe to the daily AI newsletter.
Add as Preferred Source
