About two years ago, Google practically force-fed artificial intelligence to the masses when it began showing A.I.-generated responses to people’s questions at the top of search results. Now the company is taking a similar tack by adding A.I. into another ubiquitously used service, Google Mail.
Google this month began rolling out a suite of new tools relying on generative A.I., the technology driving chatbots, to help users manage their bloated inboxes and speed up the process of writing email. Some of the features are free, while others require paying a subscription.
Gmail users can now look up emails by typing a question, such as “What’s the name of the job recruiter I met last month?” Google is also testing a new type of inbox, set for release later this year, that automatically pulls together a to-do list based on tasks discussed inside emails. In addition, Google unveiled tools to streamline writing, including an automatic proofreader and response generator.
If the new A.I. technology becomes widely adopted by Gmail users, this could be the biggest change to email, our most steadfast web service, in decades. It may transform the way people manually check their inboxes all day long into a more streamlined experience, an overview that people look at periodically.
All of this, of course, has implications for privacy. To make the new features work, Gemini, Google’s A.I. assistant, needs access to a user’s entire inbox. The company insists that while Gemini systems analyze our emails, there are protections in place so that its employees do not read them. To understand what this means for us, I interviewed a Google executive overseeing Gmail as well as privacy experts.
And to assess whether the tools are worth the potential trade-off in privacy, I tested Google’s A.I.-ified Gmail for the last week. I found a few of these features, especially the automatic to-do list generated from my inbox, very useful, but I didn’t care for the tools that sped up writing. Ultimately, I’m on the fence about whether I will opt to share even more data with Google to use these features once they are all fully released.
Here’s a rundown of first impressions of the new tools and answers to important questions about privacy going forward.
Gmail enters the A.I. era
Let’s start with the most important new feature, AI Inbox, which will be broadly released in the coming months, followed by the smaller updates available today.
A New A.I.-Powered Inbox
Managing an email inbox has become a tedious chore. When someone sends you an email, you can read and reply to it, delete it or ignore it. The option of ignoring an email takes the least effort, but unread notes pile up over time and add to stress. (Case in point: My Gmail app icon currently shows a red bubble for 131,000 unread emails.)
Google’s new AI Inbox fundamentally changes the way people check email. It gives a high-level overview of what you need to know and do today based on recent emails by looking at conversations and creating action items and summaries of topics.
The AI Inbox offered me respite from my chaotic inbox this week. I had recently exchanged emails with a local preschool discussing enrollment for my daughter. Around the same time, her pediatrician had sent an email asking me to fill out a questionnaire. But throughout the day, my inbox became cluttered with marketing messages from retailers and other unsolicited emails.
When I clicked on the AI Inbox tab, Google’s A.I. reminded me of my recent conversations and presented a to-do list: Respond to the preschool about enrollment and fill out the questionnaire for the pediatrician. None of the other noise in my inbox was there.
Google is testing the AI Inbox with a small set of users, but when the feature arrives publicly in the coming months, I think lots of people, especially busy parents and office workers, will find it useful.
Free Tools
Google is now making free the following Gmail tools that were previously available only to users who paid subscriptions to use its A.I.:
-
Suggested replies with personalization. Gemini will now analyze a message and generate a bespoke response based on your writing style in past emails.
-
AI Overviews with email summaries. Similar to the AI Overviews featured on Google search, Gmail will show an automatic summary of a conversation at the top of each email thread.
-
A “Help Me Write” button. Users can click “Help Me Write” to compose an email by typing a prompt, such as “Draft a letter to my power company asking why my bill is so high.”
As a professional writer, I mostly did not find these tools helpful, but people who struggle with words might appreciate having them.
Paid Tools
Google has also released new tools that will be available initially to people paying for a subscription to one of Google’s A.I. plans, which start at $20 a month:
-
AI Overviews for searching for emails. In the past, people could look for emails by typing keywords like “plumber” into a search bar. Now, users can type a question into the search bar, such as “What is the name of the plumber who fixed my toilet last year?”
I liked this feature — it will come in handy especially for people with bloated inboxes, since the old-school way of searching for the keyword “plumber” could load lots of irrelevant emails.
-
A proofread tool. Google’s A.I. can highlight an entire sentence that needs improvement and suggest a full revision. For example, it could trim a rambling sentence down to a few concise words.
Does Gemini in my inbox mean Google is reading my emails?
The integration of A.I. inside Gmail raises an important question about whether the technology opens doors to human reviewers reading our emails. That’s because in general, humans are involved in improving A.I. technology — for example, people occasionally need to manually review conversations with A.I. chatbots to ensure that their responses are accurate and appropriate.
In the case of Google and Gmail, the answer is, it’s complicated.
The company says that while its Gemini A.I. system can scan your emails to offer help, people do not look at your content, including the questions you ask Gemini to search for an email. In addition, the company said, it will not be using Gmail data to train, or improve, Gemini.
“We know it takes a lot of trust for people to invite A.I. to connect these dots,” said Blake Barnes, Google’s vice president of product overseeing Gmail, in an interview. “We can use A.I. to process and do these things like the AI Inbox in these isolated environments without humans in the loop at all.”
Mr. Barnes offered this analogy: Each Gmail user is living inside his or her own private room. Inside that room, Gemini is presented a question along with a set of emails that could have the answer to that question. All of this data processing happens within the private room, and the data does not leave this room to train Gemini, he added.
However, just because a company says it won’t look at your data doesn’t mean it technically can’t.
A Google spokeswoman said some user interactions with A.I. in Gmail — for example, questions posed to Gemini to search for emails — could be accessed and shared with law enforcement to comply with a legal search warrant. She noted that Google pushed back on government requests for excessive information.
This is nothing new. Google and all tech companies have been required to comply with legal requests for relevant user data, including emails and text messages, in response to search warrants and subpoenas. The only difference now is that chatting with Gemini inside Gmail is yet another source of data that can be sought through your inbox.
“It’s a reminder to people that email should be treated almost not quite public,” said Thorin Klosowski, a privacy and security activist at the Electronic Frontier Foundation, a digital rights nonprofit. “It should be treated like the company who runs it and law enforcement can access it. The more you put it in it, the more they’ll have access to.”
The upshot: If you use these tools, don’t say anything naughty to Gemini.
Should I opt out?
All of the above features, with the exception of AI Inbox, which has yet to be released, are turned on for Gmail users by default, meaning you will have to opt out if you are uninterested in using them. To opt out, go into the account settings and uncheck the box labeled “Smart features.”
Brian X. Chen is the lead consumer technology writer for The Times. He reviews products and writes Tech Fix, a column about the social implications of the tech we use.
The post A.I. Has Arrived in Gmail. Here’s What to Know. appeared first on New York Times.
