ads

Wednesday, December 23, 2020

Google reportedly tightens grip on research into ‘sensitive topics’

Google is currently under fire for apparently pushing out a researcher whose work warned of bias in AI, and now a report from Reuters says others doing such work at the company have been asked to “strike a positive tone” and undergo additional reviews for research touching on “sensitive topics.”

Reuters, citing researchers at the company and internal documents, reports that Google has implemented new controls in the last year, including an extra round of inspection for papers on certain topics and seemingly an increase in executive interference at later stages of research.

That certainly appears to have been the case with Dr. Timnit Gebru, an AI researcher at Google whose resignation seems to have been forced under confusing circumstances, following friction between her and management over work that her team was doing. (I’ve asked Gebru and Google for comment on the story.)

Among the “sensitive” topics, according to an internal webpage seen by Reuters, are: “the oil industry, China, Iran, Israel, COVID-19, home security, insurance, location data, religion, self-driving vehicles, telecoms and systems that recommend or personalize web content.”

It’s clear that many of these issues are indeed sensitive, though advising researchers to take care when addressing them seems superfluous considering the existence of ethics boards, peer review, and other ordinary controls on research. One researcher who spoke to Reuters warned that this sort of top-down interference from Google could soon get “into a serious problem of censorship.”

This is in addition to the fundamental issue of vital research being conducted under the auspices of a company for which it may or may not be in their interest to publish. Naturally large private research institutions have existed for nearly as long as organized scientific endeavor, but companies like Facebook, Google, Apple, Microsoft and others exert an enormous influence over fields like AI and have good reason to avoid criticism of lucrative technologies while shouting their usefulness from every rooftop.



from TechCrunch https://ift.tt/37LcDch
via IFTTT

No comments:

Post a Comment

Apple Vision Pro: Day One

It’s Friday, February 2, 2024. Today is the day. You’ve been eyeing the Vision Pro since Tim Cook stepped onstage with the product at last y...