Meta will track employee keystrokes and mouse movements to train AI models amid looming job cuts

Report: Meta will train AI agents by tracking employees' mouse, keyboard use

Meta will begin tracking the mouse movements, clicks, and keystrokes of its US employees to generate high-quality training data for future AI agents, Reuters reports. The news organization cites internal memos posted by the Meta Superintelligence Labs team in reporting on the new Model Capability Initiative employee-tracking software. That software will operate on specific work-related apps and websites and also make use of periodic screenshots to provide context for the AI training, according to the memo. "This is where all Meta employees can help our models get better simply by doing their daily work," the memo reads, in part, Reuters reports. Meta spokesperson Andy Stone told Reuters that the collected training data will help Meta's AI agents with tasks that it sometimes struggles with, including "things like mouse movements, clicking buttons, and navigating dropdown menus." "If we're building agents to help people complete everyday tasks using computers, our models need real examples of how we actually use them," Stone said, adding that the collected data would not be used to evaluate employees. While Meta's US employees will have their actions tracked by the new tracking software, similarly monitoring European Meta employees would likely run afoul of a number of national laws limiting how an employer can track employee actions. Meta has faced potential legal problems in the EU for forcing users of its social media services to opt-out of having their content used for AI training, rather than affirmatively opting in. Getting on the training train The Internet contains enormous amounts of text, images, and video that can be used to train generative AI models (with some important and heavily argued legal limits). But obtaining high-quality training data for physical actions or virtual computer interactions has proven more difficult. Some companies have resorted to complex physics simulations of elaborate hand-tracking prosthetics to create human interaction data that an AI robotics model can understand. Meta's move comes as major tech companies, including OpenAI, Anthropic, Google, and Perplexity, have in recent months introduced new tools that let AI agents take over your computer or web browser to complete certain tasks. Ars' own initial tests of some of these consumer offerings showed a surprising ability to convert many natural-language commands into virtual actions, with some significant limitations and brittleness regarding long-term automated tasks. Meta has also reportedly begun setting AI usage goals among some employees, including coders and engineers. The company is also reportedly planning to start laying off up to 10 percent of its global workforce starting in May.

Meta will record employees' keystrokes and use it to train its AI models | TechCrunch

Meta has found a new source of training data for its AI models: its own employees. The company plans to use data culled from the mouse movements and keystrokes of its own staff in its pursuit to build more capable and efficient artificial intelligence. The story, which was first reported by Reuters, shows the lengths to which tech companies are going to find new sources of training data -- the lifeblood of AI models that helps the programs learn how to more effectively carry out tasks and respond to user queries. When reached for comment by TechCrunch, a Meta spokesperson provided the following statement: "If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them -- things like mouse movements, clicking buttons, and navigating dropdown menus. To help, we're launching an internal tool that will capture these kinds of inputs on certain applications to help us train our models. There are safeguards in place to protect sensitive content, and the data is not used for any other purpose." This trend would seem to reveal the troublesome privacy implications of the AI industry, as yesterday's internal corporate communications are increasingly becoming fodder for a new corporate supply chain. Last week it was reported that old startups were being scavenged for their corporate communications (from Slack archives, Jira tickets and other internal messaging platforms), which could be converted into AI fuel.

Meta Is Making Workers Train Their AI Replacements

Four years ago, Mark Zuckerberg, feeling like the admiral of the good ship Meta Platforms Inc., sent a memo to employees encouraging them to refer to themselves by the nautical term "Metamates," apparently to foster a sense of camaraderie and shared mission or something. Now he plans to spy on his own Metamates to help train the robots that will someday throw them overboard. They ought to mutiny. Meta plans to put tracking software on its employees' computers to track their mouse movements, clicks and keystrokes, Reuters reported on Tuesday. It will also randomly capture images of its workers' screens every now and then. All of this is meant to help train AI models to better mimic what human beings do when they work on computers, including tricky stuff like using drop-down menus. One reason to do such a thing is to help eventually replace those humans with AI models, which don't require paychecks or health insurance or common courtesy. In Zuckerberg's February 2022 memo, he suggested his employees adopt a motto of "Meta, Metamates, Me." This borrowed from a US Navy saying, "Ship, shipmates, self." In the Navy's context, it means that, when your ship is on fire or its toilets don't work or some other emergency, your first priority is taking care of the ship. Your shipmates are your next concern. Only after they're safe can you think about yourself. This hierarchy makes sense in a military context, where one willingly signs away one's life to be a weapon for one's country. It makes less sense in the context of writing computer code for a company that began as a website for ranking women by hotness and has since transitioned to other pursuits such as democracy-poisoning, virtual reality and, lately, artificial intelligence. Meta's tracking, known as the Model Capability Initiative, will apply only to work-related apps and sites, according to the internal memo Reuters obtained. So supposedly it won't teach robots how to order from DoorDash or Amazon, wasteful things humans often do instead of working. Nor will it be used for performance evaluations or other human-relations purposes, a Meta spokesperson reassured everybody. Its sole purpose is to "help our models get better simply ⁠by doing their daily work," according to the memo. But, even if you set aside panopticon-related concerns, this is still not good news for Metamates. The MCI will be just part of a broader data-harvesting campaign in service of AI for Work, Meta's self-described "effort to integrate AI into every tool, team, and process at Meta." That will necessarily mean fewer human workers; by May 20, in fact, there will be about 8,000 fewer of those, according to an earlier Reuters report. Meta plans to lay off that many Metamates, or about 10% of its total workforce, on that date. And that will just be the first round of job cuts. The second, of what could be a similar size, will come later in the year. Meta's expensive investment in AI is partly to blame. The company recently projected capital spending of $115 billion to $135 billion this year, more than double 2025's spree, most of which is going toward AI. Human paychecks must be sacrificed to help cover the costs. But Zuckerberg also clearly expects robots to replace those employees, telling Joe Rogan last year that he expected to soon have "an AI that can effectively be a sort of mid-level engineer that you have at your company that can write code." If you are currently a mid-level engineer at Meta, this can't be reassuring. Knowing that your keystrokes and mouse clicks are helping to hasten your replacement, then perhaps you will be motivated to keystroke and mouse-click much less productively. This also speaks to the broader zeitgeist, in which 55% of Americans think AI will do them more harm than good, including taking their jobs. They have little faith it will be of much benefit to anyone but the CEOs like Zuckerberg who are spending hundreds of billions of dollars to make it happen. Meta's stock price has gone basically nowhere for the past two years. Investors may be waiting to see if Zuckerberg's massive investment in AI goes any better than his massive, failed investment in the Metaverse (remember that?). If his robots don't live up to his hopes, then he may well find he has fewer willing and able hands on deck to steer him to his next misadventure. More From Bloomberg Opinion: * Meta's Victory Lap Over Google on Ads Will Be Short: Dave Lee * Doomscrolling Dangers Are a Worthy Legal Target: Parmy Olson * A Fix for Gambling, Emissions and AI's Ills: Nicolas Rohatyn Want more Bloomberg Opinion? Terminal readers, head to OPIN <GO>. Or subscribe to our daily newsletter.

Exclusive: Meta to start capturing employee mouse movements, keystrokes for AI training data

NEW YORK, April 21 (Reuters) - Meta (META.O), opens new tab is installing new tracking software on U.S.-based employees' computers to capture mouse movements, clicks and keystrokes for use in training its artificial-intelligence models, part of a broad initiative to build AI agents that can perform work tasks autonomously, the company told staffers in internal memos seen by Reuters. The tool will run on a list of work-related apps and websites and will also take ⁠occasional snapshots of the content on employees' screens for context, according to one memo, posted by a staff AI research scientist on Tuesday in a dedicated internal channel for the company's model-building Meta SuperIntelligence Labs team. The purpose of the exercise, according to the memo, was to improve the company's models in areas where they still struggle, like choosing from dropdown menus and using keyboard shortcuts. "This is where all Meta employees ⁠can help our models get better simply by doing their daily work," it said. Meta spokesperson Andy Stone said the data collected would not be used for performance assessments or any other purpose besides model training and that ⁠safeguards were in place to protect sensitive content. "If we're building agents to help people complete everyday tasks using computers, our models need real examples of how ⁠people actually use them -- things like mouse movements, clicking buttons, and navigating dropdown menus. To help, we're launching an internal tool that will ⁠capture these kinds of inputs on certain applications to help us train our models," said Stone. Reporting by Katie Paul in New York and Jeff Horwitz in San Francisco; Editing by Matthew Lewis Contact: [email protected] Our Standards: The Thomson Reuters Trust Principles., opens new tab

Meta to track workers' clicks and keystrokes to train AI

Meta will start tracking the way employees work, including their keystrokes and mouse clicks, to train its artificial intelligence (AI) models. The company, which owns Instagram and Facebook, told workers on Tuesday that a new tool will run on Meta's computers and internal apps, logging their activity to be used as training data for AI technology. A Meta spokesman told the BBC: "If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them." "The data is not used for any other purpose," he said, adding that the tool has "safeguards in place to protect sensitive content". But one Meta employee, who asked not to be identified, said having their smallest actions on a computer being used to train AI model as workers expect a slew of additional job cuts feels "very dystopian". "This company has become obsessed with AI," they told the BBC.. Another person who recently left the company said the tracking tool is "just the latest way they're shoving AI down everyone's throat". Meta has already laid off around 2,000 employees this year in smaller rounds of cuts, but employees have been expecting deeper job losses in the coming months, as the BBC previously reported. Last month, the company enacted a partial hiring freeze which now appears to be more far-reaching. A website that Meta uses to advertise all of its jobs hosted about 800 listings in March. Now, it is advertising just seven jobs. Meta spokesman's declined to comment on the company's removal of job listings or plans for cuts. Meta is using a tracking tool called Model Capability Initiative or MCI, according to Reuters which first reported the move. The BBC has been told that an employee's activity on a Meta computer would have been accessible to the company before, however tracking and logging specifically for the purpose of training and improving AI tools is new. Mark Zuckerberg, Meta's co-founder and chief executive, recently pledged to ramp up spending on AI projects this year and is attempting to position the firm at the forefront of the technology. Meta plans to spend roughly $140bn on AI in 2026, almost double the amount it invested in the technology a year ago. In 2025, it effectively acquired Scale AI for more than $14bn (£10.3bn), and brought the executives of the data-labeling firm into Meta to help it build out its AI models and tools. The first significant launch from the company's reformed Meta Superintelligence Labs group emerged last month with the AI model Muse Spark. With the data gathered from the new employee tracker, Meta is hoping to train new AI models that will come out of the lab. In January, Zuckerberg said that 2026 will be "the year that AI dramatically changes the way we work". "We're starting to see projects that used to take big teams now be accomplished by a single, very talented person," he said at the time.

Meta Plans to Turn Its Employees' Clicks and Keystrokes into AI Training Data

Much has been made about the idea that workers are effectively training their own replacements when they work with AI tools, though most employers won't directly admit to it. Meta has apparently decided to drop all pretenses that isn't the case. According to a report from Reuters, the company recently sent a memo to employees informing them of new tracking software that will be installed on their computers to track mouse movements and keystrokes in order to help train AI agents to perform specific work tasks. Per the memo seen by Reuters, the surveillance tool (no reason to pretend it's anything but that) is called Model Capability Initiative, and it will record the screens of employees as they go about their work. In addition, the company will reportedly increase its internal data collection efforts as part of its AI for Work program, which has apparently been renamed the Agent Transformation Accelerator. All of that data will be used to train Meta's AI models to improve the functionality of its agents, which are meant to operate autonomously and navigate across multiple systems and programs. Per Reuters, Meta's current agent offerings reportedly struggle to perform certain actions that humans have no problem with, like selecting items from dropdown menus or using keyboard shortcuts. The new employee monitoring systems will help to refine those capabilities, which definitely should not raise any red flags for the employees of a company that is reportedly gearing up to lay off 10% of its workforce in the coming months. The pitch to employees is that their work won't change as a result of having persistent digital eyes on them. The company positioned the program as an opportunity for Meta employees to "help our models get better simply by doing their daily work," with a promise that the information would not be used for performance reviews or other potentially invasive purposes. It's hard to imagine that's a compelling case for employees, who don't really have any incentive to participate in such a program. They've added model training to their responsibilities without any apparent increase in compensation, and the new work requires them to surrender a sense of privacy that their work isn't being monitored. Meta is clearly all in on AI. The company has reportedly been building AI agents meant to work alongside employees, including one designed for CEO Mark Zuckerberg, and has created a Zuckerberg chatbot for employees to communicate with. Now workers are essentially being told they are training the systems that will replace them. Might be time to figure out how to poison the well.

Meta Installing Software on Employee Computers to Track Everything They Do, Feed the Data to AI

Can't-miss innovations from the bleeding edge of science and tech As if activity-monitoring software installed on your work computer that snitches on you if you're away from the keyboard for too long wasn't enough, Meta is taking the trend to its logical -- and dystopian surveillance state-level -- conclusion. As Reuters reports, the Mark Zuckerberg-led company is installing new tracking software on all of its US-based employees' computers that tracks all of their mouse movements and keystrokes, data that will be used for training the company's AI models. The company is reportedly looking to develop AI agents that can complete work tasks autonomously, in perhaps one of the more conspicuous efforts to automate human workers's jobs we've come across as of late. Besides the ethical concerns of forcing employees to train their AI replacements, the news also raises thorny questions regarding data privacy. Meta, in particular, has garnered an incredibly poor reputation when it comes to protecting personal data. According to an internal memo obtained by Reuters, the software is called "Model Capability Initiative" and will run on work-related apps and websites. It will even take occasional screenshots. The goal is to guide Meta's AI models to essentially replicate the way humans interact with computers, like using dropdown menus or making use of keyboard shortcuts. "This is where all Meta employees can help our models get better simply by doing their daily work," the memo reads, as quoted by Reuters. Meta spokesperson Andy Stone told Reuters that the company is looking to implement safeguards to protect "sensitive content," but didn't elaborate any further. While tracking employees' keystrokes and mouse movement would likely be against European law, Yale University law professor Ifeoma Ajunwa told Reuters that "there is no limit on worker surveillance" in the US on a federal level. Beyond tracking their employees' every move, Meta is also planning to slash ten percent of its workforce across the globe starting next month -- only the first of several planned cuts later this year.

Meta will start tracking employees' screens and keystrokes to train AI tools | Fortune

The tool, disclosed in a memo to staff this week in a channel belonging to the Meta Superintelligence Labs team, which Reuters saw, will run on a designated list of work apps and websites. Per Reuters, the memo framed the effort as a way for rank-and-file employees to improve company models in areas where they struggle to emulate basic computer-use behaviors, such as navigating dropdown menus and using keyboard shortcuts. The memo told Meta staffers that they can do their part to help by just doing their daily work. The broader goal seems to be to build AI agents capable of performing white-collar tasks on their own, the exact software Meta is racing to ship out amid competition from OpenAI and Anthropic. Those agents have a lot of data, but little footage of how to actually use it. "If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them," a Meta spokesperson wrote in an email to Fortune, adding that the models were using "things like mouse movements, clicking buttons, and navigating dropdown menus." The company added that safeguards are in place to protect sensitive content and that the data will not be used for any other purpose. The move comes as the industry hunts for training data to use in the workplace itself. In January, OpenAI was reported to be asking third-party contractors, via training data firm Handshake AI, to upload samples of real work products from previous jobs -- actual PowerPoints, spreadsheets, and the like -- with instructions to scrub confidential material before submission. Meta acquired a 49% stake in data-labeling firm Scale AI last year for more than $14 billion, and Scale's former CEO, Alexandr Wang, now leads Meta Superintelligence Labs. Meta has rapidly accelerated its AI spending, with CEO Mark Zuckerberg committing up to $135 billion in capital expenditure for 2026. At the same time, the company is preparing to cut as much as 20% of its workforce, with the first layoffs reportedly set to begin in May.

Would you quit? Meta will put keyloggers on employee PCs for AI training

Meta is also planning more layoffs as it spends billions on AI. It's best to assume that nothing you do on a work-issued PC is private. But is there no limit? No breaking point where people start throwing their laptops out of windows and joining Watch Dogs-esque hacktivist groups? Meta sure seems to be looking for it. Reuters reports that the Facebook, VR, and now AI company will track US-based employees' "mouse movements, clicks and keystrokes" on "work-related apps and websites" for the purpose of AI training. It will also take screenshots, according to the report. A Meta spokesperson told Reuters that the company will somehow exclude "sensitive content" and won't use the data for performance evaluations. Rather, their AI models "need real examples" of people using computers for everyday tasks in order to automate them. If I were a Meta employee, I'd take this to mean that the company thinks I can be replaced by a robot that I was forced to train. Meta has pledged to spend $600 billion on AI by 2028, and according to another Reuters report, plans to lay off nearly 8,000 employees in May. The mood inside Meta has been "horrid" in recent years, according to Ed Zitron, an AI industry critic who reports on what he's called "the most annoying bubble in history" in his Where's Your Ed At newsletter. Zitron's impression from sources inside the company is that there exists a "culture of paranoia," which an AI-training keylogger won't likely improve. "Everyone I know at Meta hates working there," Zitron told PC Gamer on a call today. Meta introduced a new AI model, Muse Spark, earlier in April, and claims that it's a step toward "superintelligence," a buzzword the AI industry has been bandying about for years now. A Meta executive acknowledged to Bloomberg that the model performs worse than competitor models at some tasks, but said that it's "early," and that the company has bigger LLMs in development. In its report, Reuters notes that Meta's keylogging initiative likely targets US employees because laws related to employee surveillance are stricter in Europe.

Meta to start capturing employee mouse movements, keystrokes for AI training data - The Economic Times

Meta is installing new tracking software on US-based employees' computers to capture mouse movements, clicks and keystrokes for use in training its artificial-intelligence models, part of a broad initiative to build AI agents that can perform work tasks autonomously, ⁠the ⁠company told staffers in internal memos seen by Reuters. The tool will run on a list of work-related apps and websites and will also take occasional snapshots of the content on employees' screens for context, according to one memo, posted by a staff AI research scientist on Tuesday in a dedicated ⁠internal channel for the company's model-building Meta SuperIntelligence Labs team. The purpose of the exercise, according to the memo, was ⁠to improve the company's models in areas where they still struggle, like choosing from dropdown menus and using keyboard shortcuts. "This is where all Meta employees can help our models get better simply by doing their daily work," it said. Meta spokesperson Andy Stone said the data collected would not be used for performance assessments or any other purpose besides model training and that safeguards were in place to protect sensitive content. "If we're building agents to help people ⁠complete everyday tasks using computers, our models need real examples of how people actually use them - things like mouse movements, clicking buttons, and navigating dropdown menus. To help, we're launching an internal tool that will capture these kinds of inputs on certain applications to help us train our models," said Stone.

Meta Plans To Log Employee Keystrokes For Training Workplace AI Agents: Report - Broadcom (NASDAQ:AVGO),

How Meta Intends To Turn Daily Work Into AI Training Reuters reported that Meta is rolling out a tool called Model Capability Initiative, or MCI, that runs across work apps and websites while recording clicks, cursor movement, and typing. One internal memo said the system can also capture periodic screenshots, with the goal of improving areas where models still stumble on routine computer behavior such as navigating dropdowns or using keyboard shortcuts. The memo framed the effort as a way for staff to contribute training examples simply by doing their normal jobs, rather than creating separate labeling workflows. In a separate message shared on Monday, Meta CTO Andrew Bosworth described a broader internal program now labeled Agent Transformation Accelerator, built around AI agents doing more of the work while employees supervise and refine outcomes. Safeguards, Limits, And A Growing Surveillance Debate Meta spokesperson Andy Stone said the MCI data is intended only for model training and not for employee evaluation, while also pointing to protections meant to keep sensitive information out of the pipeline. As reported by Reuters, Yale law professor Ifeoma Ajunwa warned that logging keystrokes can push white-collar monitoring toward a level of continuous oversight more commonly associated with gig-economy roles. Outside the U.S., legal constraints could be tougher, according to Reuters, which cited York University professor Valerio De Stefano on how European privacy and labor rules may restrict or bar this kind of tracking. De Stefano also said the mere knowledge of monitoring can tilt workplace leverage toward employers, even before any data is used. Why Custom Chips Matter For Meta's Agent Ambitions Meta's internal data collection drive is paired with a hardware strategy built for long-run AI capacity, including a multi-year expansion with Broadcom to design custom silicon tailored to Meta's workloads. That kind of contracted, purpose-built chip pipeline is a departure from buying only off-the-shelf parts, and it can make AI spending look more like an ongoing infrastructure commitment than a series of one-time orders. The same dynamic has been pitched as a stabilizer for chip suppliers, since long-duration agreements can provide visibility that historically was hard to find in boom-and-bust semiconductor cycles. META Price Action: Meta Platforms shares were down 0.31% during regular trading and up 0.45% in after-hours trading on Tuesday, last trading at $671.85, according to Benzinga Pro data. Image: Shutterstock This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.

Meta to track employee keystrokes to train AI models, Reuters reports By Investing.com

Investing.com -- Meta Platforms (NASDAQ:META) is installing tracking software on U.S.-based employees' computers to capture mouse movements, clicks and keystrokes for artificial intelligence training data, according to internal memos seen by Reuters. The tool, called Model Capability Initiative (MCI), will monitor a list of work-related apps and websites and take occasional snapshots of screen content, according to a memo posted Tuesday by a staff AI research scientist in a channel for the company's Meta SuperIntelligence Labs team. The company said the data will be used to train AI models to perform work tasks autonomously, focusing on areas where the technology struggles to replicate human computer interactions, such as selecting from dropdown menus and using keyboard shortcuts. "This is where all Meta employees can help our models get better simply by doing their daily work," the memo said. The announcement follows a Monday memo from Meta CTO Andrew Bosworth, who told employees the company would increase internal data collection as part of its AI for Work (AI4W) efforts. This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.