From Gmail to Word, your privacy settings and AI are entering into a new relationship
The Microsoft 365 website on a laptop arranged in New York, US, on Tuesday, June 25, 2024.
Bloomberg | Bloomberg | Getty Images
The beginning of the year is a great time to do some basic cyber hygiene. We’ve all been told to patch, change passwords, and update software. But one concern that has been increasingly creeping to the forefront is the sometimes quiet integration of potentially privacy-invading AI into programs.
“AI’s rapid integration into our software and services has and should continue to raise significant questions about privacy policies that preceded the AI era,” said Lynette Owens, vice president, global consumer education at cybersecurity company Trend Micro. Many programs we use today — whether it be email, bookkeeping, or productivity tools, and social media and streaming apps — may be governed by privacy policies that lack clarity on whether our personal data can be used to train AI models.
“This leaves all of us vulnerable to uses of our personal information without the appropriate consent. It’s time for every app, website, or online service to take a good hard look at the data they are collecting, who they’re sharing it with, how they’re sharing it, and whether or not it can be accessed to train AI models,” Owens said. “There’s a lot of catch up needed to be done.”
Where AI is already inside our daily online lives
Owens said the potential issues overlap with most of the programs and applications we use on a daily basis.
“Many platforms have been integrating AI into their operations for years, long before AI became a buzzword,” she said.
As an example, Owens points out that Gmail has used AI for spam filtering and predictive text with its “Smart Compose” feature. “And streaming services like Netflix rely on AI to analyze viewing habits and recommend content,” Owens said. Social media platforms like Facebook and Instagram have long used AI for facial recognition in photos and personalized content feeds.
“While these tools offer convenience, consumers should consider the potential privacy trade-offs, such as how much personal data is being collected and how it is used to train AI systems. Everyone should carefully review privacy settings, understand what data is being shared, and regularly check for updates to terms of service,” Owens said.
One tool that has come in for particular scrutiny is Microsoft’s connected experiences, which has been around since 2019 and comes activated with an optional opt-out. It was recently highlighted in press reports — inaccurately, according to the company as well as some outside cybersecurity experts that have taken a look at the issue — as a feature that is new or that has had its settings changed. Leaving the sensational headlines aside, privacy experts do worry that advances in AI can lead to the potential for data and words in programs like Microsoft Word to be used in ways that privacy settings do not adequately cover.
“When tools like connected experiences evolve, even if the underlying privacy settings haven’t changed, the implications of data use might be far broader,” Owens said.
A spokesman for Microsoft wrote in a statement to CNBC that Microsoft does not use customer data from Microsoft 365 consumer and commercial applications to train foundational large language models. He added that in certain instances, customers may consent to using their data for specific purposes, such as custom model development explicitly requested by some commercial customers. Additionally, the setting enables cloud-backed features many people have come to expect from productivity tools such as real-time co-authoring, cloud storage and tools like Editor in Word that provide spelling and grammar suggestions.
Default privacy settings are an issue
Ted Miracco, CEO of security software company Approov, said features like Microsoft’s connected experiences are a double-edged sword — the promise of enhanced productivity but the introduction of significant privacy red flags. The setting’s default-on status could, Miracco said, opt people into something they aren’t necessarily aware of, primarily related to data collection, and organizations may also want to think twice before leaving the feature on.
“Microsoft’s assurance provides only partial relief, but still falls short of mitigating some real privacy concern,” Miracco said.
Perception can be its own problem, according to Kaveh Vadat, founder of RiseOpp, an SEO marketing agency.
“Having the default to enablement shifts the dynamic significantly,” Vahdat said. “Automatically enabling these features, even with good intentions, inherently places the onus on users to review and modify their privacy settings, which can feel intrusive or manipulative to some.”
His view is that companies need to be more transparent, not less, in an environment where there is a lot of distrust and suspicion regarding AI.
Companies including Microsoft should emphasize default opt-out rather than opt-in, and might provide more granular, non-technical information about how personal content is handled because perception can become a reality.
“Even if the technology is completely safe, public perception is shaped not just by facts but by fears and assumptions — especially in the AI era where users often feel disempowered,” he said.
Default settings that enable sharing make sense for business reasons but are bad for consumer privacy, according to Jochem Hummel, assistant professor of information systems and management at Warwick Business School at the University of Warwick in England.
Companies are able to enhance their products and maintain competitiveness with more data sharing as the default, Hummel said. However, from a user standpoint, prioritizing privacy by adopting an opt-in model for data sharing would be “a more ethical approach,” he said. And as long as the additional features offered through data collection are not indispensable, users can choose which aligns more closely with their interests.
There are real benefits to the current tradeoff between AI-enhanced tools and privacy, Hummel said, based on what he is seeing in the work turned in by students. Students who have grown up with web cameras, lives broadcast in real-time on social media, and all-encompassing technology, are often less concerned about privacy, Hummel said, and are embracing these tools enthusiastically. “My students, for example, are creating better presentations than ever,” he said.
Managing the risks
In areas such as copyright law, fears about massive copying by LLMs have been overblown, according to Kevin Smith, director of libraries at Colby College, but AI’s evolution does intersect with core privacy concerns.
“A lot of the privacy concerns currently being raised about AI have actually been around for years; the rapid deployment of large language model trained AI has just focused attention on some of those issues,” Smith said. “Personal information is all about relationships, so the risk that AI models could uncover data that was more secure in a more ‘static’ system is the real change we need to find ways to manage,” he added.
In most programs, turning off AI features is an option buried in the settings. For instance, with connected experiences, open a document and then click “file” and then go to “account” and then find privacy settings. Once there, go to “manage settings” and scroll down to connected experiences. Click the box to turn it off. Once doing so, Microsoft warns: “If you turn this off, some experiences may not be available to you.” Microsoft says leaving the setting on will allow for more communication, collaboration, and AI served-up suggestions.
In Gmail, one needs to open it, tap the menu, then go to settings, then click the account you want to change and then scroll to the “general” section and uncheck the boxes next to the various “Smart features” and personalization options.
As cybersecurity vendor Malwarebytes put it in a blog post about the Microsoft feature: “turning that option off might result in some lost functionality if you’re working on the same document with other people in your organization. … If you want to turn these settings off for reasons of privacy and you don’t use them much anyway, by all means, do so. The settings can all be found under Privacy Settings for a reason. But nowhere could I find any indication that these connected experiences were used to train AI models.”
While these instructions are easy enough to follow, and learning more about what you have agreed to is probably a good option, some experts say the onus should not be on the consumer to deactivate these settings. “When companies implement features like these, they often present them as opt-ins for enhanced functionality, but users may not fully understand the scope of what they’re agreeing to,” said Wes Chaar, a data privacy expert.
“The crux of the issue lies in the vague disclosures and lack of clear communication about what ‘connected’ entails and how deeply their personal content is analyzed or stored,” Chaar said. “For those outside of technology, it might be likened to inviting a helpful assistant into your home, only to learn later they’ve taken notes on your private conversations for a training manual.”
The decision to manage, limit, or even revoke access to data underscores the imbalance in the current digital ecosystem. “Without robust systems prioritizing user consent and offering control, individuals are left vulnerable to having their data repurposed in ways they neither anticipate nor benefit from,” Chaar said.
Technology
AI research takes a backseat to profits as Silicon Valley prioritizes products over safety, experts say
Sam Altman, co-founder and CEO of OpenAI and co-founder of Tools for Humanity, participates remotely in a discussion on the sidelines of the IMF/World Bank Spring Meetings in Washington, D.C., April 24, 2025.
Brendan Smialowski | AFP | Getty Images
Not long ago, Silicon Valley was where the world’s leading artificial intelligence experts went to perform cutting-edge research.
Meta, Google and OpenAI opened their wallets for top talent, giving researchers staff, computing power and plenty of flexibility. With the support of their employers, the researchers published high-quality academic papers, openly sharing their breakthroughs with peers in academia and at rival companies.
But that era has ended. Now, experts say, AI is all about the product.
Since OpenAI released ChatGPT in late 2022, the tech industry has shifted its focus to building consumer-ready AI services, in many cases prioritizing commercialization over research, AI researchers and experts in the field told CNBC. The profit potential is massive — some analysts predict $1 trillion in annual revenue by 2028. The prospective repercussions terrify the corner of the AI universe concerned about safety, industry experts said, particularly as leading players pursue artificial general intelligence, or AGI, which is technology that rivals or exceeds human intelligence.
In the race to stay competitive, tech companies are taking an increasing number of shortcuts when it comes to the rigorous safety testing of their AI models before they are released to the public, industry experts told CNBC.
James White, chief technology officer at cybersecurity startup CalypsoAI, said newer models are sacrificing security for quality, that is, better responses by the AI chatbots. That means they’re less likely to reject malicious kinds of prompts that could cause them to reveal ways to build bombs or sensitive information that hackers could exploit, White said.
“The models are getting better, but they’re also more likely to be good at bad stuff,” said White, whose company performs safety and security audits of popular models from Meta, Google, OpenAI and other companies. “It’s easier to trick them to do bad stuff.”
The changes are readily apparent at Meta and Alphabet, which have deprioritized their AI research labs, experts say. At Facebook’s parent company, the Fundamental Artificial Intelligence Research, or FAIR, unit has been sidelined by Meta GenAI, according to current and former employees. And at Alphabet, the research group Google Brain is now part of DeepMind, the division that leads development of AI products at the tech company.
CNBC spoke with more than a dozen AI professionals in Silicon Valley who collectively tell the story of a dramatic shift in the industry away from research and toward revenue-generating products. Some are former employees at the companies with direct knowledge of what they say is the prioritization of building new AI products at the expense of research and safety checks. They say employees face intensifying development timelines, reinforcing the idea that they can’t afford to fall behind when it comes to getting new models and products to market. Some of the people asked not to be named because they weren’t authorized to speak publicly on the matter.
Mark Zuckerberg, CEO of Meta Platforms, during the Meta Connect event in Menlo Park, California, on Sept. 25, 2024.
David Paul Morris | Bloomberg | Getty Images
Meta’s AI evolution
When Joelle Pineau, a Meta vice president and the head of the company’s FAIR division, announced in April that she would be leaving her post, many former employees said they weren’t surprised. They said they viewed it as solidifying the company’s move away from AI research and toward prioritizing developing practical products.
“Today, as the world undergoes significant change, as the race for AI accelerates, and as Meta prepares for its next chapter, it is time to create space for others to pursue the work,” Pineau wrote on LinkedIn, adding that she will formally leave the company May 30.
Pineau began leading FAIR in 2023. The unit was established a decade earlier to work on difficult computer science problems typically tackled by academia. Yann LeCun, one of the godfathers of modern AI, initially oversaw the project, and instilled the research methodologies he learned from his time at the pioneering AT&T Bell Laboratories, according to several former employees at Meta. Small research teams could work on a variety of bleeding-edge projects that may or may not pan out.
The shift began when Meta laid off 21,000 employees, or nearly a quarter of its workforce, starting in late 2022. CEO Mark Zuckerberg kicked off 2023 by calling it the “year of efficiency.” FAIR researchers, as part of the cost-cutting measures, were directed to work more closely with product teams, several former employees said.
Two months before Pineau’s announcement, one of FAIR’s directors, Kim Hazelwood, left the company, two people familiar with the matter said. Hazelwood helped oversee FAIR’s NextSys unit, which manages computing resources for FAIR researchers. Her role was eliminated as part of Meta’s plan to cut 5% of its workforce, the people said.
Joelle Pineau of Meta speaks at the Advancing Sustainable Development through Safe, Secure, and Trustworthy AI event at Grand Central Terminal in New York, Sept. 23, 2024.
Bryan R. Smith | Via Reuters
OpenAI’s 2022 launch of ChatGPT caught Meta off guard, creating a sense of urgency to pour more resources into large language models, or LLMs, that were captivating the tech industry, the people said.
In 2023, Meta began heavily pushing its freely available and open-source Llama family of AI models to compete with OpenAI, Google and others.
With Zuckerberg and other executives convinced that LLMs were game-changing technologies, management had less incentive to let FAIR researchers work on far-flung projects, several former employees said. That meant deprioritizing research that could be viewed as having no impact on Meta’s core business, such as FAIR’s previous health care-related research into using AI to improve drug therapies.
Since 2024, Meta Chief Product Officer Chris Cox has been overseeing FAIR as a way to bridge the gap between research and the product-focused GenAI group, people familiar with the matter said. The GenAI unit oversees the Llama family of AI models and the Meta AI digital assistant, the two most important pillars of Meta’s AI strategy.
Under Cox, the GenAI unit has been siphoning more computing resources and team members from FAIR due to its elevated status at Meta, the people said. Many researchers have transferred to GenAI or left the company entirely to launch their own research-focused startups or join rivals, several of the former employees said.
While Zuckerberg has some internal support for pushing the GenAI group to rapidly develop real-world products, there’s also concern among some staffers that Meta is now less able to develop industry-leading breakthroughs that can be derived from experimental work, former employees said. That leaves Meta to chase its rivals.
A high-profile example landed in January, when Chinese lab DeepSeek released its R1 model, catching Meta off guard. The startup claimed it was able to develop a model as capable as its American counterparts but with training at a fraction of the cost.
Meta quickly implemented some of DeepSeek’s innovative techniques for its Llama 4 family of AI models that were released in April, former employees said. The AI research community had a mixed reaction to the smaller versions of Llama 4, but Meta said the biggest and most powerful Llama 4 variant is still being trained.
The company in April also released security and safety tools for developers to use when building apps with Meta’s Llama 4 AI models. These tools help mitigate the chances of Llama 4 unintentionally leaking sensitive information or producing harmful content, Meta said.
“Our commitment to FAIR remains strong,” a Meta spokesperson told CNBC. “Our strategy and plans will not change as a result of recent developments.”
In a statement to CNBC, Pineau said she is enthusiastic about Meta’s overall AI work and strategy.
“There continues to be strong support for exploratory research and FAIR as a distinct organization in Meta,” Pineau said. “The time was simply right for me personally to re-focus my energy before jumping into a new adventure.”
Meta on Thursday named FAIR co-founder Rob Fergus as Pineau’s replacement. Fergus will return to the company to serve as a director at Meta and head of FAIR, according to his LinkedIn profile. He was most recently a research director at Google DeepMind.
“Meta’s commitment to FAIR and long term research remains unwavering,” Fergus said in a LinkedIn post. “We’re working towards building human-level experiences that transform the way we interact with technology and are dedicated to leading and advancing AI research.”
Demis Hassabis, co-founder and CEO of Google DeepMind, attends the Artificial Intelligence Action Summit at the Grand Palais in Paris, Feb. 10, 2025.
Benoit Tessier | Reuters
Google ‘can’t keep building nanny products’
Google released its latest and most powerful AI model, Gemini 2.5, in March. The company described it as “our most intelligent AI model,” and wrote in a March 25 blog post that its new models are “capable of reasoning through their thoughts before responding, resulting in enhanced performance and improved accuracy.”
For weeks, Gemini 2.5 was missing a model card, meaning Google did not share information about how the AI model worked or its limitations and potential dangers upon its release.
Model cards are a common tool for AI transparency.
A Google website compares model cards to food nutrition labels: They outline “the key facts about a model in a clear, digestible format,” the website says.
“By making this information easy to access, model cards support responsible AI development and the adoption of robust, industry-wide standards for broad transparency and evaluation practices,” the website says.
Google wrote in an April 2 blog post that it evaluates its “most advanced models, such as Gemini, for potential dangerous capabilities prior to their release.” Google later updated the blog to remove the words “prior to their release.”
Without a model card for Gemini 2.5, the public had no way of knowing which safety evaluations were conducted or whether DeepMind checked for dangerous capabilities at all.
In response to CNBC’s inquiry on April 2 about Gemini 2.5’s missing model card, a Google spokesperson said that a “tech report with additional safety information and model cards are forthcoming.” Google published an incomplete model card on April 16 and updated it on April 28, more than a month after the AI model’s release, to include information about Gemini 2.5’s “dangerous capability evaluations.”
Those assessments are important for gauging the safety of a model — whether people can use the models to learn how to build chemical or nuclear weapons or hack into important systems. These checks also determine whether a model is capable of autonomously replicating itself, which could lead to a company losing control of it. Running tests for those capabilities requires more time and resources than simple, automated safety evaluations, according to industry experts.
Google co-founder Sergey Brin
Kelly Sullivan | Getty Images Entertainment | Getty Images
The Financial Times in March reported that Google DeepMind CEO Demis Hassabis had installed a more rigorous vetting process for internal research papers to be published. The clampdown at Google is particularly notable because the company’s “Transformers” technology gained recognition across Silicon Valley through that type of shared research. Transformers were critical to OpenAI’s development of ChatGPT and the rise of generative AI.
Google co-founder Sergey Brin told staffers at DeepMind and Gemini in February that competition has accelerated and “the final race to AGI is afoot,” according to a memo viewed by CNBC. “We have all the ingredients to win this race but we are going to have to turbocharge our efforts,” he said in the memo.
Brin said in the memo that Google has to speed up the process of testing AI models, as the company needs “lots of ideas that we can test quickly.”
“We need real wins that scale,” Brin wrote.
In his memo, Brin also wrote that the company’s methods have “a habit of minor tweaking and overfitting” products for evaluations and “sniping” the products at checkpoints. He said employees need to build “capable products” and to “trust our users” more.
“We can’t keep building nanny products,” Brin wrote. “Our products are overrun with filters and punts of various kinds.”
A Google spokesperson told CNBC that the company has always been committed to advancing AI responsibly.
“We continue to do that through the safe development and deployment of our technology, and research contributions to the broader ecosystem,” the spokesperson said.
Sam Altman, CEO of OpenAI, is seen through glass during an event on the sidelines of the Artificial Intelligence Action Summit in Paris, Feb. 11, 2025.
Aurelien Morissard | Via Reuters
OpenAI’s rush through safety testing
The debate of product versus research is at the center of OpenAI’s existence. The company was founded as a nonprofit research lab in 2015 and is now in the midst of a contentious effort to transform into a for-profit entity.
That’s the direction co-founder and CEO Sam Altman has been pushing toward for years. On May 5, though, OpenAI bowed to pressure from civic leaders and former employees, announcing that its nonprofit would retain control of the company even as it restructures into a public benefit corporation.
Nisan Stiennon worked at OpenAI from 2018 to 2020 and was among a group of former employees urging California and Delaware not to approve OpenAI’s restructuring effort. “OpenAI may one day build technology that could get us all killed,” Stiennon wrote in a statement in April. “It is to OpenAI’s credit that it’s controlled by a nonprofit with a duty to humanity.”
But even with the nonprofit maintaining control and majority ownership, OpenAI is speedily working to commercialize products as competition heats up in generative AI. And it may have rushed the rollout of its o1 reasoning model last year, according to some portions of its model card.
Results of the model’s “preparedness evaluations,” the tests OpenAI runs to assess an AI model’s dangerous capabilities and other risks, were based on earlier versions of o1. They had not been run on the final version of the model, according to its model card, which is publicly available.
Johannes Heidecke, OpenAI’s head of safety systems, told CNBC in an interview that the company ran its preparedness evaluations on near-final versions of the o1 model. Minor variations to the model that took place after those tests wouldn’t have contributed to significant jumps in its intelligence or reasoning and thus wouldn’t require additional evaluations, he said. Still, Heidecke acknowledged that OpenAI missed an opportunity to more clearly explain the difference.
OpenAI’s newest reasoning model, o3, released in April, seems to hallucinate more than twice as often as o1, according to the model card. When an AI model hallucinates, it produces falsehoods or illogical information.
OpenAI has also been criticized for reportedly slashing safety testing times from months to days and for omitting the requirement to safety test fine-tuned models in its latest “Preparedness Framework.”
Heidecke said OpenAI has decreased the time needed for safety testing because the company has improved its testing effectiveness and efficiency. A company spokesperson said OpenAI has allocated more AI infrastructure and personnel to its safety testing, and has increased resources for paying experts and growing its network of external testers.
In April, the company shipped GPT-4.1, one of its new models, without a safety report, as the model was not designated by OpenAI as a “frontier model,” which is a term used by the tech industry to refer to a bleeding-edge, large-scale AI model.
But one of those small revisions caused a big wave in April. Within days of updating its GPT-4o model, OpenAI rolled back the changes after screenshots of overly flattering responses to ChatGPT users went viral online. OpenAI said in a blog post explaining its decision that those types of responses to user inquiries “raise safety concerns — including around issues like mental health, emotional over-reliance, or risky behavior.”
OpenAI said in the blogpost that it opted to release the model even after some expert testers flagged that its behavior “‘felt’ slightly off.”
“In the end, we decided to launch the model due to the positive signals from the users who tried out the model. Unfortunately, this was the wrong call,” OpenAI wrote. “Looking back, the qualitative assessments were hinting at something important, and we should’ve paid closer attention. They were picking up on a blind spot in our other evals and metrics.”
Metr, a company OpenAI partners with to test and evaluate its models for safety, said in a recent blog post that it was given less time to test the o3 and o4-mini models than predecessors.
“Limitations in this evaluation prevent us from making robust capability assessments,” Metr wrote, adding that the tests it did were “conducted in a relatively short time.”
Metr also wrote that it had insufficient access to data that would be important in determining the potential dangers of the two models.
The company said it wasn’t able to access the OpenAI models’ internal reasoning, which is “likely to contain important information for interpreting our results.” However, Metr said, “OpenAI shared helpful information on some of their own evaluation results.”
OpenAI’s spokesperson said the company is piloting secure ways of sharing chains of thought for Metr’s research as well as for other third-party organizations.
Steven Adler, a former safety researcher at OpenAI, told CNBC that safety testing a model before it’s rolled out is no longer enough to safeguard against potential dangers.
“You need to be vigilant before and during training to reduce the chance of creating a very capable, misaligned model in the first place,” Adler said.
He warned that companies such as OpenAI are backed into a corner when they create capable but misaligned models with goals that are different from the ones they intended to build.
“Unfortunately, we don’t yet have strong scientific knowledge for fixing these models — just ways of papering over the behavior,” Adler said.
WATCH: OpenAI closes $40 billion funding round, largest private tech deal on record
Omar Marques | Sopa Images | Lightrocket | Getty Images
Shares of stock brokerage platform eToro popped in their Nasdaq debut on Wednesday after the company raised almost $310 million in its initial public offering.
The stock opened at $69.69, or 34% above its IPO, pushing its market cap to $5.6 billion. Shares were last up more than 40%.
The Israel-based company sold nearly six million shares at $52 each, above the expected range of $46 to $50. Almost six million additional shares were sold by existing investors. At the IPO price, the company was valued at roughly $4.2 billion.
Wall Street is looking to the Robinhood competitor for signs of renewed interest in IPOs after an extended drought. Many investors saw President Donald Trump’s return to the White House as a catalyst before tariff concerns led companies to delay their plans.
Etoro isn’t the only company attempting to test the waters. Fintech company Chime filed its prospectus with the U.S. Securities and Exchange Commission on Tuesday, while digital physical therapy company Hinge Health kickstarted its IPO roadshow, and said in a filing it aims to raise up to $437 million in its offering.
EToro had previously filed to go public in 2021 through a merger with a special purpose acquisition company, or SPAC, that would have valued it at more than $10 billion. It shelved those plans in 2022 as equity markets nosedived, but remained focused on an eventual IPO.
EToro was founded in 2007 by brothers Yoni and Ronen Assia and David Ring. The company makes money through trading-related fees and nontrading activities such as withdrawals. Net income increased almost thirteenfold last year to $192.4 million from $15.3 million in 2023.
The company has steadily built a growing business in cryptocurrencies. Revenue from crypto assets more than tripled to upward of $12 million in 2024, and one-quarter of its net trading contribution stemmed from crypto last year. That is up from 10% in 2023.
EToro said that for the first quarter, it expects crypto assets to account for 37% of its commission from trading activities, down from 43% a year earlier.
Spark Capital is the company’s biggest outside investor, with 14% control after the offering, followed by BRM Group at 8.7%. CEO Yoni Assia controls 9.3%.
-
Sports3 years ago‘Storybook stuff’: Inside the night Bryce Harper sent the Phillies to the World Series
-
Sports1 year agoStory injured on diving stop, exits Red Sox game
-
Sports2 years agoGame 1 of WS least-watched in recorded history
-
Sports2 years agoMLB Rank 2023: Ranking baseball’s top 100 players
-
Sports4 years ago
Team Europe easily wins 4th straight Laver Cup
-
Environment2 years agoJapan and South Korea have a lot at stake in a free and open South China Sea
-
Environment2 years agoGame-changing Lectric XPedition launched as affordable electric cargo bike
-
Business3 years agoBank of England’s extraordinary response to government policy is almost unthinkable | Ed Conway