Google AI Creeps Into Gmail Accounts to ‘Find Exactly What You’re Looking For’

Google AI Creeps Into Gmail Accounts to ‘Find Exactly What You’re Looking For’
The Google logo is pictured atop an office building in Irvine, Calif., on Aug. 7, 2017. (Mike Blake/Reuters)
Naveen Athrappully
6/4/2023
Updated:
6/8/2023

Google is introducing a “highly requested” AI search feature in its Gmail service amid concerns that artificial intelligence could pose a threat to privacy.

The new Gmail AI search feature is intended to “improve the search experience” of Gmail mobile users and will help people find “exactly what you’re looking for,” Google said in a June 2 blog post. “When searching in Gmail, machine learning models will use the search term, most recent emails, and other relevant factors to show you the results that best match your search query. These results will now appear at the top of the list in a dedicated section, followed by all results sorted by recency.”

“This highly requested feature gives you the most relevant information first, allowing you to more quickly and easily find specific emails or files.” The Gmail AI search update began rolling out on June 2.

Google’s decision to allow AI to sort through people’s emails raises privacy concerns. Such issues were earlier brought to attention in a March 21 post on twitter by author Kate Crawford.

When she asked Google’s Bard AI where its dataset comes from, the artificial intelligence replied that one of the sources is Google’s internal data, including data from Gmail.

“I’m assuming that’s flat-out wrong; otherwise, Google is crossing some serious legal boundaries,” Crawford said at the time. Google later replied to the tweet that Bard was “not trained on Gmail data.”

The problem with AI having access to personal data is that artificial intelligence could end up manipulating human beings.

In an April 25 interview with The Epoch Times, Rep. Jay Obernolte (R-Calif.) pointed out that AIs have an “uncanny ability to pierce through personal digital privacy” that could help governments predict and control people’s behavior.
“I worry about the way that AI can empower a nation-state to create, essentially, a surveillance state, which is what China is doing with it,” Obernolte said.

AI Manipulating People

The manipulative abilities of AI could be used by businesses to boost consumerist tendencies among people. In a June 2 interview with The Epoch Times, music expert Peter Tregear said that AI-generated music, which is already spreading through social media, ads, and shopping centers, can encourage shoppers to buy more.

“It will be so much easier and cheaper to underscore visual material that it becomes ubiquitous … You see people walking around and basically wired in 24/7. They wake up to music, put in their headphones, and have their phone all day. Once they take it out, they’re in a shop which has music in the background,” he said.

“You’re not aware of it. That’s what makes it so manipulative because it ‘disappears,’ it literally becomes an underscore.”

AI can also be used to track and manipulate voters. Ethan Busby, a political psychologist, told The Epoch Times that AI has read so many things produced by human beings that they can “say back to us things about ourselves that we didn’t necessarily know.”

“You can basically ask these tools to put themselves in a specific frame of mind and pretend to be essentially this person, pretend to have these characteristics.”

In a study, Busby and his team at Brigham Young University looked at using AI to predict voting patterns. Analyzing the 2016 elections, the team found that the AI quickly learned to predict how people would vote based on analyzing their attributes.

Daniel. Y. Teng contributed to this report.