Google

AI Overviews gets upgraded to Gemini 3 with a dash of AI Mode

AI, ai overviews, Artificial Intelligence, generative ai, Google, search / Mike M. / January 27, 2026

It can be hard sometimes to keep up with the deluge of generative AI in Google products. Even if you try to avoid it all, there are some features that still manage to get in your face. Case in point: AI Overviews. This AI-powered search experience has a reputation for getting things wrong, but you may notice some improvements soon. Google says AI Overviews is being upgraded to the latest Gemini 3 models with a more conversational bent.

In just the last year, Google has radically expanded the number of searches on which you get an AI Overview at the top. Today, the chatbot will almost always have an answer for your query, which has relied mostly on models in Google’s Gemini 2.5 family. There was nothing wrong with Gemini 2.5 as generative AI models go, but Gemini 3 is a little better by every metric.

There are, of course, multiple versions of Gemini 3, and Google doesn’t like to be specific about which ones appear in your searches. What Google does say is that AI Overviews chooses the right model for the job. So if you’re searching for something simple for which there are a lot of valid sources, AI Overviews may manifest something like Gemini 3 Flash without running through a ton of reasoning tokens. For a complex “long tail” query, it could step up the thinking or move to Gemini 3 Pro (for paying subscribers).

AI Overviews gets upgraded to Gemini 3 with a dash of AI Mode Read More »

“Wildly irresponsible”: DOT’s use of AI to draft safety rules sparks concerns

AI, AI hallucination, Artificial Intelligence, department of transportation, Gemini, Google, Policy / Mike M. / January 26, 2026

At DOT, Trump likely hopes to see many rules quickly updated to modernize airways and roadways. In a report highlighting the Office of Science and Technology Policy’s biggest “wins” in 2025, the White House credited DOT with “replacing decades-old rules with flexible, innovation-friendly frameworks,” including fast-tracking rules to allow for more automated vehicles on the roads.

Right now, DOT expects that Gemini can be relied on to “handle 80 to 90 percent of the work of writing regulations,” ProPublica reported. Eventually all federal workers who rely on AI tools like Gemini to draft rules “would fall back into merely an oversight role, monitoring ‘AI-to-AI interactions,’” ProPublica reported.

Google silent on AI drafting safety rules

Google did not respond to Ars’ request to comment on this use case for Gemini, which could spread across government under Trump’s direction.

Instead, the tech giant posted a blog on Monday, pitching Gemini for government more broadly, promising federal workers that AI would help with “creative problem-solving to the most critical aspects of their work.”

Google has been competing with AI rivals for government contracts, undercutting OpenAI and Anthropic’s $1 deals by offering a year of access to Gemini for $0.47.

The DOT contract seems important to Google. In a December blog, the company celebrated that DOT was “the first cabinet-level agency to fully transition its workforce away from legacy providers to Google Workspace with Gemini.”

At that time, Google suggested this move would help DOT “ensure the United States has the safest, most efficient, and modern transportation system in the world.”

Immediately, Google encouraged other federal leaders to launch their own efforts using Gemini.

“We are committed to supporting the DOT’s digital transformation and stand ready to help other federal leaders across the government adopt this blueprint for their own mission successes,” Google’s blog said.

DOT did not immediately respond to Ars’ request for comment.

“Wildly irresponsible”: DOT’s use of AI to draft safety rules sparks concerns Read More »

Google begins offering free SAT practice tests powered by Gemini

AI, Artificial Intelligence, education, Google / Paul Patrick / January 22, 2026

It’s no secret that students worldwide use AI chatbots to do their homework and avoid learning things. On the flip side, students can also use AI as a tool to beef up their knowledge and plan for the future with flashcards or study guides. Google hopes its latest Gemini feature will help with the latter. The company has announced that Gemini can now create free SAT practice tests and coach students to help them get higher scores.

As a standardized test, the content of the SAT follows a predictable pattern. So there’s no need to use a lengthy, personalized prompt to get Gemini going. Just say something like, “I want to take a practice SAT test,” and the chatbot will generate one complete with clickable buttons, graphs, and score analysis.

Of course, generative AI can go off the rails and provide incorrect information, which is a problem when you’re trying to learn things. However, Google says it has worked with education firms like The Princeton Review to ensure the AI-generated tests resemble what students will see in the real deal.

The interface for Gemini’s practice tests includes scoring and the ability to review previous answers. If you are unclear on why a particular answer is right or wrong, the questions have an “Explain answer” button right at the bottom. After you finish the practice exam, the custom interface (which looks a bit like Gemini’s Canvas coding tool) can help you follow up on areas that need improvement.

Google begins offering free SAT practice tests powered by Gemini Read More »

Google adds your Gmail and Photos to AI Mode to enable “Personal Intelligence”

AI, Artificial Intelligence, Gemini, generative ai, Google, search / Tim Belzer / January 22, 2026

Google believes AI is the future of search, and it’s not shy about saying it. After adding account-level personalization to Gemini earlier this month, it’s now updating AI Mode with so-called “Personal Intelligence.” According to Google, this makes the bot’s answers more useful because they are tailored to your personal context.

Starting today, the feature is rolling out to all users who subscribe to Google AI Pro or AI Ultra. However, it will be a Labs feature that needs to be explicitly enabled (subscribers will be prompted to do this). Google tends to expand access to new AI features to free accounts later on, so free users will most likely get access to Personal Intelligence in the future. Whenever this option does land on your account, it’s entirely optional and can be disabled at any time.

If you decide to integrate your data with AI Mode, the search bot will be able to scan your Gmail and Google Photos. That’s less extensive than the Gemini app version, which supports Gmail, Photos, Search, and YouTube history. Gmail will probably be the biggest contributor to AI Mode—a great many life events involve confirmation emails. Traditional search results when you are logged in are adjusted based on your usage history, but this goes a step further.

If you’re going to use AI Mode to find information, Personal Intelligence could actually be quite helpful. When you connect data from other Google apps, Google’s custom Gemini search model will instantly know about your preferences and background—that’s the kind of information you’d otherwise have to include in your search query to get the best output. With Personal Intelligence, AI Mode can just pull those details from your email or photos.

Google adds your Gmail and Photos to AI Mode to enable “Personal Intelligence” Read More »

Google temporarily disabled YouTube’s advanced captions without warning

Google, Tech, youtube / Kris Guyer / January 20, 2026

YouTubers have been increasingly frustrated with Google’s management of the platform, with disinformation welcomed back and an aggressive push for more AI (except where Google doesn’t like it). So it’s no surprise that creators have been up in arms over the suspicious removal of YouTube’s advanced SRV3 caption format. You don’t have to worry too much just yet—Google says this is only temporary, and it’s working on a fix for the underlying bug.

Google added support for this custom subtitle format around 2018, giving creators more customization options than with traditional captions. SRV3 (also known as YTT or YouTube Timed Text) allows for custom colors, transparency, animations, fonts, and precise positioning in videos. Uploaders using this format can color-code and position captions to help separate multiple speakers, create sing-along animations, or style them to match the video.

Over the last several days, creators who’ve become accustomed to this level of control have been dismayed to see that YouTube is no longer accepting videos with this Google-created format. Many worried Google had ditched the format entirely, which could be problematic for all those previously uploaded videos.

Google has now posted a brief statement and confirmed to Ars that it has not ended support for SRV3. However, all is not well. The company says it has temporarily limited the serving of SRV3 caption files because they may break playback for some users. That’s pretty vague, but it sounds like developers made a change to the platform without taking into account how it might interfere with SRV3 captions. Rather than allow those videos to be non-functional, it’s disabling most of the captions.

Google temporarily disabled YouTube’s advanced captions without warning Read More »

OpenAI to test ads in ChatGPT as it burns through billions

advertising, AI, AI assistants, AI bubble, Biz & IT, chatbots, chatgpt, chatgtp, Fidji Simo, generative ai, Google, machine learning, openai, sam altman / Tim Belzer / January 16, 2026

Financial pressures and a changing tune

OpenAI’s advertising experiment reflects the enormous financial pressures facing the company. OpenAI does not expect to be profitable until 2030 and has committed to spend about $1.4 trillion on massive data centers and chips for AI.

According to financial documents obtained by The Wall Street Journal in November, OpenAI expects to burn through roughly $9 billion this year while generating $13 billion in revenue. Only about 5 percent of ChatGPT’s 800 million weekly users pay for subscriptions, so it’s not enough to cover all of OpenAI’s operating costs.

Not everyone is convinced ads will solve OpenAI’s financial problems. “I am extremely bearish on this ads product,” tech critic Ed Zitron wrote on Bluesky. “Even if this becomes a good business line, OpenAI’s services cost too much for it to matter!”

OpenAI’s embrace of ads appears to come reluctantly, since it runs counter to a “personal bias” against advertising that Altman has shared in earlier public statements. For example, during a fireside chat at Harvard University in 2024, Altman said he found the combination of ads and AI “uniquely unsettling,” implying that he would not like it if the chatbot itself changed its responses due to advertising pressure. He added: “When I think of like GPT writing me a response, if I had to go figure out exactly how much was who paying here to influence what I’m being shown, I don’t think I would like that.”

An example mock-up of an advertisement in ChatGPT provided by OpenAI. Credit: OpenAI

Along those lines, OpenAI’s approach appears to be a compromise between needing ad revenue and not wanting sponsored content to appear directly within ChatGPT’s written responses. By placing banner ads at the bottom of answers separated from the conversation history, OpenAI appears to be addressing Altman’s concern: The AI assistant’s actual output, the company says, will remain uninfluenced by advertisers.

Indeed, Simo wrote in a blog post that OpenAI’s ads will not influence ChatGPT’s conversational responses and that the company will not share conversations with advertisers and will not show ads on sensitive topics such as mental health and politics to users it determines to be under 18.

“As we introduce ads, it’s crucial we preserve what makes ChatGPT valuable in the first place,” Simo wrote. “That means you need to trust that ChatGPT’s responses are driven by what’s objectively useful, never by advertising.”

OpenAI to test ads in ChatGPT as it burns through billions Read More »

TSMC says AI demand is “endless” after record Q4 earnings

AI, AI chips, AI infrastructure, Amazon, arizona, Biz & IT, C.C. Wei, datacenters, Google, machine learning, microsoft, NVIDIA, semiconductors, Taiwan, TSMC / Tim Belzer / January 16, 2026

TSMC posted net income of NT$505.7 billion (about $16 billion) for the quarter, up 35 percent year over year and above analyst expectations. Revenue hit $33.7 billion, a 25.5 percent increase from the same period last year. The company expects nearly 30 percent revenue growth in 2026 and plans to spend between $52 billion and $56 billion on capital expenditures this year, up from $40.9 billion in 2025.

Checking with the customers’ customers

Wei’s optimism stands in contrast to months of speculation about whether the AI industry is in a bubble. In November, Google CEO Sundar Pichai warned of “irrationality” in the AI market and said no company would be immune if a potential bubble bursts. OpenAI’s Sam Altman acknowledged in August that investors are “overexcited” and that “someone” will lose a “phenomenal amount of money.”

But TSMC, which manufactures the chips that power the AI boom, is betting the opposite way, with Wei telling analysts he spoke directly to cloud providers to verify that demand is real before committing to the spending increase.

“I want to make sure that my customers’ demand are real. So I talked to those cloud service providers, all of them,” Wei said. “The answer is that I’m quite satisfied with the answer. Actually, they show me the evidence that the AI really helps their business.”

The earnings report landed the same day the US and Taiwan finalized a trade agreement that cuts tariffs on Taiwanese goods to 15 percent, down from 20 percent. The deal commits Taiwanese companies to $250 billion in direct US investment, and TSMC is accelerating the expansion of its Arizona chip fabrication facilities to match.

TSMC says AI demand is “endless” after record Q4 earnings Read More »

Many Bluetooth devices with Google Fast Pair vulnerable to “WhisperPair” hack

bluetooth, Google, Security, Tech / Paul Patrick / January 15, 2026

Pairing Bluetooth devices can be a pain, but Google Fast Pair makes it almost seamless. Unfortunately, it may also leave your headphones vulnerable to remote hacking. A team of security researchers from Belgium’s KU Leuven University has revealed a vulnerability dubbed WhisperPair that allows an attacker to hijack Fast Pair-enabled devices to spy on the owner.

Fast Pair is widely used, and your device may be vulnerable even if you’ve never used a Google product. The bug affects more than a dozen devices from 10 manufacturers, including Sony, Nothing, JBL, OnePlus, and Google itself. Google has acknowledged the flaw and notified its partners of the danger, but it’s up to these individual companies to create patches for their accessories. A full list of vulnerable devices is available on the project’s website.

The researchers say that it takes only a moment to gain control of a vulnerable Fast Pair device (a median of just 10 seconds) at ranges up to 14 meters. That’s near the limit of the Bluetooth protocol and far enough that the target wouldn’t notice anyone skulking around while they hack headphones.

Once an attacker has forced a connection to a vulnerable audio device, they can perform relatively innocuous actions, such as interrupting the audio stream or playing audio of their choice. However, WhisperPair also allows for location tracking and microphone access. So the attacker can listen in on your conversations and follow you around via the Bluetooth device in your pocket. The researchers have created a helpful video dramatization (below) that shows how WhisperPair can be used to spy on unsuspecting people.

Many Bluetooth devices with Google Fast Pair vulnerable to “WhisperPair” hack Read More »

Hegseth wants to integrate Musk’s Grok AI into military networks this month

agentic AI, AI, AI image generator, AI in education, AI regulation, AI security, Anthropic, Biz & IT, censorship, Deepfakes, Department of Defense, elon musk, Gemini, Google, grok, machine learning, openai, Pentagon, Pete Hegseth, Policy, xAI / Paul Patrick / January 13, 2026

On Monday, US Defense Secretary Pete Hegseth said he plans to integrate Elon Musk’s AI tool, Grok, into Pentagon networks later this month. During remarks at the SpaceX headquarters in Texas reported by The Guardian, Hegseth said the integration would place “the world’s leading AI models on every unclassified and classified network throughout our department.”

The announcement comes weeks after Grok drew international backlash for generating sexualized images of women and children, although the Department of Defense has not released official documentation confirming Hegseth’s announced timeline or implementation details.

During the same appearance, Hegseth rolled out what he called an “AI acceleration strategy” for the Department of Defense. The strategy, he said, will “unleash experimentation, eliminate bureaucratic barriers, focus on investments, and demonstrate the execution approach needed to ensure we lead in military AI and that it grows more dominant into the future.”

As part of the plan, Hegseth directed the DOD’s Chief Digital and Artificial Intelligence Office to use its full authority to enforce department data policies, making information available across all IT systems for AI applications.

“AI is only as good as the data that it receives, and we’re going to make sure that it’s there,” Hegseth said.

If implemented, Grok would join other AI models the Pentagon has adopted in recent months. In July 2025, the defense department issued contracts worth up to $200 million for each of four companies, including Anthropic, Google, OpenAI, and xAI, for developing AI agent systems across different military operations. In December 2025, the Department of Defense selected Google’s Gemini as the foundation for GenAI.mil, an internal AI platform for military use.

Hegseth wants to integrate Musk’s Grok AI into military networks this month Read More »

Microsoft vows to cover full power costs for energy-hungry AI data centers

AI, AI infrastructure, AI training, Amazon, Biz & IT, brad smith, community development, datacenters, Donald Trump, electricity, Google, machine learning, Meta, microsoft, openai, property taxes, skilled trades, water use / Paul Patrick / January 13, 2026

Taking responsibility for power usage

In the Microsoft blog post, Smith acknowledged that residential electricity rates have recently risen in dozens of states, driven partly by inflation, supply chain constraints, and grid upgrades. He wrote that communities “value new jobs and property tax revenue, but not if they come with higher power bills or tighter water supplies.”

Microsoft says it will ask utilities and public commissions to set rates high enough to cover the full electricity costs for its data centers, including infrastructure additions. In Wisconsin, the company is supporting a new rate structure that would charge “Very Large Customers,” including data centers, the cost of the electricity required to serve them.

Smith wrote that while some have suggested the public should help pay for the added electricity needed for AI, Microsoft disagrees. He stated, “Especially when tech companies are so profitable, we believe that it’s both unfair and politically unrealistic for our industry to ask the public to shoulder added electricity costs for AI.”

On water usage for cooling, Microsoft plans a 40 percent improvement in data center water-use intensity by 2030. A recent environmental audit from AI model-maker Mistral found that training and running its Large 2 model over 18 months produced 20.4 kilotons of CO₂ emissions and evaporated enough water to fill 112 Olympic-size swimming pools, illustrating the aggregate environmental impact of AI operations at scale.

To solve some of these issues, Microsoft says it has launched a new AI data center design using a closed-loop system that constantly recirculates cooling liquid, dramatically cutting water usage. In this design, already deployed in Wisconsin and Georgia, potable water is no longer needed for cooling.

On property taxes, Smith stated in the blog post that the company will not ask local municipalities to reduce their rates. The company says it will pay its full share of local property taxes. Smith wrote that Microsoft’s goal is to bring these commitments to life in the first half of 2026. Of course, these are PR-aligned company goals and not realities yet, so we’ll have to check back in later to see whether Microsoft has been following through on its promises.

Microsoft vows to cover full power costs for energy-hungry AI data centers Read More »

Google’s updated Veo model can make vertical videos from reference images with 4K upscaling

AI, Ai video, generative ai, Google, Google Veo / Mike M. / January 13, 2026

Enhanced support for Ingredients to Video and the associated vertical outputs are live in the Gemini app today, as well as in YouTube Shorts and the YouTube Create app, fulfilling a promise initially made last summer. Veo videos are short—just eight seconds long for each prompt. It would be tedious to assemble those into a longer video, but Veo is perfect for the Shorts format.

Veo 3.1 Updates – Seamlessly blend textures, characters, and objects.

The new Veo 3.1 update also adds an option for higher-resolution video. The model now supports 1080p and 4K outputs. Google debuted 1080p support last year, but it’s mentioning that option again today, suggesting there may be some quality difference. 4K support is new, but neither 1080p nor 4K outputs are native. Veo creates everything in 720p resolution, but it can be upscaled “for high-fidelity production workflows,” according to Google. However, a Google rep tells Ars that upscaling is only available in Flow, the Gemini API, and Vertex AI. Video in the Gemini app is always 720p.

We are rushing into a world where AI video is essentially indistinguishable from real life. Google, which more or less controls online video via YouTube’s dominance, is at the forefront of that change. Today’s update is reasonably significant, and it didn’t even warrant a version number change. Perhaps we can expect more 2025-style leaps in video quality this year, for better or worse.

Google’s updated Veo model can make vertical videos from reference images with 4K upscaling Read More »

Google removes some AI health summaries after investigation finds “dangerous” flaws

AI, AI hallucination, ai overviews, Biz & IT, generative ai, Google, health, machine learning, misinformation, search engines, the guardian / Paul Patrick / January 12, 2026

Why AI Overviews produces errors

The recurring problems with AI Overviews stem from a design flaw in how the system works. As we reported in May 2024, Google built AI Overviews to show information backed up by top web results from its page ranking system. The company designed the feature this way based on the assumption that highly ranked pages contain accurate information.

However, Google’s page ranking algorithm has long struggled with SEO-gamed content and spam. The system now feeds these unreliable results to its AI model, which then summarizes them with an authoritative tone that can mislead users. Even when the AI draws from accurate sources, the language model can still draw incorrect conclusions from the data, producing flawed summaries of otherwise reliable information.

The technology does not inherently provide factual accuracy. Instead, it reflects whatever inaccuracies exist on the websites Google’s algorithm ranks highly, presenting the facts with an authority that makes errors appear trustworthy.

Other examples remain active

The Guardian found that typing slight variations of the original queries into Google, such as “lft reference range” or “lft test reference range,” still prompted AI Overviews. Hebditch said this was a big worry and that the AI Overviews present a list of tests in bold, making it very easy for readers to miss that these numbers might not even be the right ones for their test.

AI Overviews still appear for other examples that The Guardian originally highlighted to Google. When asked why these AI Overviews had not also been removed, Google said they linked to well-known and reputable sources and informed people when it was important to seek out expert advice.

Google said AI Overviews only appear for queries where it has high confidence in the quality of the responses. The company constantly measures and reviews the quality of its summaries across many different categories of information, it added.

This is not the first controversy for AI Overviews. The feature has previously told people to put glue on pizza and eat rocks. It has proven unpopular enough that users have discovered that inserting curse words into search queries disables AI Overviews entirely.

Google removes some AI health summaries after investigation finds “dangerous” flaws Read More »