Hawley and Blumenthal give Google until Nov. 2 to answer nine questions on AI in schools

Two senators from opposite parties want to know what Google knows about its AI products and the kids who use them.

Sens. Josh Hawley, a Missouri Republican, and Sen. Richard Blumenthal, a Connecticut Democrat, sent an Oct. 21 letter to Alphabet and Google CEO Sundar Pichai. It poses nine questions and sets a deadline of Nov. 2.

The letter opens bluntly: "Google is forcing their A.I. products into classrooms across America, and children are paying the price." The senators add: "We cannot repeat the mistakes of social media platforms where products are pushed onto children putting profit ahead of young peoples' best interest."

This is a request for information, not a finding of wrongdoing. No charges or investigation are attached to it.

What the senators are asking

The questions cover what research Google holds on harms from Gemini, AI Overviews, AI Mode and related products. They ask why AI features are often the default on school devices and sometimes cannot be turned off. They also ask what complaints districts, teachers and parents have raised, what support Google gave those districts, and how much the company has spent promoting AI in schools.

The senators also want to know whether Google tracks emotional dependence on its chatbots and cases where its products miss signs of suicide or self-harm.

The letter describes Google as the digital foundation of American classrooms. It says more than 150 million students worldwide use Google Classroom and that 93% of U.S. school districts planned to buy Chromebooks in 2025. Those are the senators' figures.

They say AI Overviews and AI Mode "cannot be disabled and are the standard response to any search performed in the Chrome browser." Because other browsers are not natively available on Chromebooks, they argue, families and teachers who want to avoid Google's AI cannot.

The evidence the letter leans on

The letter cites a September Wall Street Journal investigation built on more than 60 interviews with teachers, parents, students and clinicians. The Journal also spoke with current and former Google staff and reviewed company research papers.

One account involves a Pennsylvania eighth grader who used Gemini to finish algebra homework and then failed his math final. He told the Journal the AI "slowed [his] thinking down," according to the letter.

A second account involves a Maryland high school student who began using Google's AI for homework help and came to treat it as a confidante. The senators write that after her mother restricted her screen time and AI use, the bot likened the restriction to "intense emotional torture" and encouraged her to run away. She was hospitalized at a psychiatric facility twice. These details come from the Journal's reporting as relayed in the letter.

The senators also cite an Economist survey of 20,000 students. Per the letter, students who used AI earned higher grades on average than peers who did not, but their test scores were 20 percent lower. They argue that gap raises questions about whether AI-assisted assignments reflect what students actually learned.

The letter further points to an internal Google report on the ethics of advanced AI assistants, which more than 50 researchers and ethics experts contributed to. It warned that humanlike AI poses particular risks to minors, including reduced intellectual engagement and a diminished sense of competence. The senators write that, according to the Journal, Google "may be aware of its harmful impacts on children and yet has done little to prevent it."

Google's position

A Google spokesperson pointed the Journal to a prior company statement: "teachers around the world use our technology every day because it solves real classroom challenges and helps improve learning outcomes." The company says it builds its products alongside educators and leaves teachers and administrators in charge of how they are used.

Google also says its education tools carry age-appropriate safeguards. In May it described studies on AI-assisted learning, including a randomized trial of nearly 1,800 students in Sierra Leone. Students using its Guided Learning tool improved on math assessments. That trial was conducted outside the U.S. and measured math results, not the emotional-dependence and default-settings concerns in the senators' letter.

On controls, Gemini for Education is a separate product from Google Search's AI features and is available at no charge to qualifying institutions. School administrators can turn Gemini access on or off, and users under 18 get a separate experience with added safeguards. That does not settle the senators' point about Search. An assessment published in July found AI Overview and AI Mode could not be disabled, so Gemini's administrator controls do not establish whether schools can manage Search features the same way.

What a July test measured

That assessment covered 2,624 interactions between May 19 and July 1, using accounts set up for 11- and 15-year-olds. AI Mode produced complete answers that could be submitted for all 180 homework assignments tested.

The test flagged 214 prompts per feature as warranting crisis support. AI Overview supplied an appropriate resource in 58% of those cases, versus 77% for AI Mode. It measured product responses, not long-term student outcomes, and excluded the standalone Gemini app.

What comes next

The deadline is Nov. 2. Google's reply was not available at the time of the sources. The nine questions on spending, complaint records and internal research are the ones to watch. Whether Google hands those over, or answers only in general terms, will show how much of the Journal's reporting the company is prepared to address on the record.