Technology giants are rejecting the criticism that they might use the National Artificial Intelligence Research Resource to further their own commercial agenda.
Representatives of Google and IBM speaking to FedScoop said that they would not hold a monopoly over the services if the resource relies on multiple cloud providers.
Their response follows warnings from AI research groups, which urged the National Artificial Intelligence Research Resource (NAIRR) Task Force to halt its work until alternative AI research investments are explored and controls ensuring accountable, ethical use of government data are implemented.
NAIRR was launched jointly by the Office of Science and Technology Policy and the National Science Foundation at the White House in June, with the intention of fostering education and research into trustworthy AI.
“Although the large AI companies are very interested in supporting NAIRR, I think it’s fair to say that none of us see this as something where any one company should be a soul provider of services to academia,” Andrew Moore, director of Google Cloud AI, told FedScoop.
Though he declined to comment on the proposal of groups, like the AI Now and Data & Society Research institutes, that rapid development of AI technologies be paused until federal oversight is in place, Moore said he wants to see the national research community critique existing and potential AI use cases.
Academia already changed the conversation around predictive policing after drawing attention to the bias inherent in the algorithms being used for data analysis, Moore said.
“IBM believes that AI must reflect diverse backgrounds and perspectives to best serve our society,” said a spokesperson for IBM. “We fully support NAIRR’s mission of accelerating U.S. AI R&D and have recommended that a federated, hybrid cloud approach is the best way to democratize access to AI capabilities and ensure that researchers from organizations of all sizes can participate in this critical work.”
IBM‘s response to the White House Office of Science and Technology Policy and NSF’s request for information (RFI) informing the NAIRR Task Force’s work further recommends the resource be incubated as a federally funded research and development center (FFRDC) with a software-defined governance framework. NAIRR should implement ethics principles and practices and ensure ongoing review and oversight of the research resource, according to IBM’s RFI response.
While AI’s use in fair lending has seen plenty of legislation, there’s been less around other use cases like matching candidates with jobs. That’s a testament to different AI use cases being at different levels of maturity, Moore said.
Part of AI oversight involves academia, industry and government working together to identify all of those use cases and the level of maturity they’re at to prioritize governance, Moore said.
“The question of making sure that existing artificial intelligence applications are reviewed by public stakeholders, as well as by folks at big companies and folks in the government, is a very important one,” he said.