Google tells its scientists to be more positive about AI

Google
(Image credit: Shutterstock / rvlsoft)

Google has asked its researchers to “strike a positive tone” and subject their research on “sensitive topics”, such as Artificial Intelligence (AI), to additional reviews, new reports have claimed.

Asking for an extra round of inspection is supposedly part of new controls implemented by the search giant for papers on certain topics. 

If true, they aren’t going to win Google any brownie points, which drew a lot of flak earlier this month for allegedly firing Dr. Timnit Gebru, the co-lead of Google’s Ethical AI team, after supposed disagreements with her work.

Concern or censorship?

Reports now cite researchers and internal documents that show that in at least three cases senior managers have requested authors to refrain from casting its technology in a negative light.

The report quotes one such document as saying “Advances in technology and the growing complexity of our external environment are increasingly leading to situations where seemingly inoffensive projects raise ethical, reputational, regulatory or legal issues.” 

In another alleged internal correspondence, a senior Google manager reviewing a study on content recommendation technology shortly before publication is quoted as asking authors to “take great care to strike a positive tone.” 

The leaked documents, whose authenticity can’t be verified by TechRadar Pro, spell out some of the “sensitive” topics as “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.”

The reports also quote senior Google AI researcher Margaret Mitchell as saying that this sort of interference could be seen as a form of censorship: 

“If we are researching the appropriate thing given our expertise, and we are not permitted to publish that on grounds that are not in line with high-quality peer review, then we’re getting into a serious problem of censorship.”

Via: Reuters

TOPICS
Mayank Sharma

With almost two decades of writing and reporting on Linux, Mayank Sharma would like everyone to think he’s TechRadar Pro’s expert on the topic. Of course, he’s just as interested in other computing topics, particularly cybersecurity, cloud, containers, and coding.

Latest in Pro
Half man, half AI.
Three key AI considerations for engineering leaders
Vodafone logo outside a store in Sydney
Vodafone employees could lose bonuses if they’re not in office 8 days per month
Homepage of Manus, a new Chinese artificial intelligence agent capable of handling complex, real-world tasks, is seen on the screen of an iPhone.
Manus AI may be the new DeepSeek, but initial users report problems
healthcare
Software bug meant NHS information was potentially “vulnerable to hackers”
Hospital
Major Oracle outage hits US Federal health record systems
A hacker wearing a hoodie sitting at a computer, his face hidden.
Experts warn this critical PHP vulnerability could be set to become a global problem
Latest in News
Apple's Craig Federighi demonstrates the iPhone Mirroring feature of macOS Sequoia at the Worldwide Developers Conference (WWDC) 2024.
Report: iOS 19 and macOS 16 could mark their biggest design overhaul in years – and we have one request
Google Gemini Calendar
Gemini is coming to Google Calendar, here’s how it will work and how to try it now
Lego Mario Kart – Mario & Standard Kart set on a shelf.
Lego just celebrated Mario Day in the best way possible, with an incredible Mario Kart set that's up for preorder now
TCL QM7K TV on orange background
TCL’s big, bright new mid-range mini-LED TVs have built-in Bang & Olufsen sound
Apple iPhone 16e
Which affordable phone wins the mid-range race: the iPhone 16e, Nothing 3a, or Samsung Galaxy A56? Our latest podcast tells all
An image of a Jackbox Games Party Pack
Jackbox games is coming to smart TVs in mid-2025, and I can’t wait to be reunited with one of my favorite party video games