Meta pauses employee tracker for AI training amid privacy concerns
Mark Zuckerberg’s Meta has paused a program that tracked employees’ computer activity amid data privacy concerns and a staff backlash.
The owner of Facebook, Instagram and WhatsApp had introduced a tool that tracked staff keystrokes, mouse clicks and content displayed on computer screens in order to collect data for training its AI models.
More than 1,600 Meta workers signed a petition against the tool, called the Model Capability Initiative, demanding the company does not harvest “employee ‘computer use’ data”.
The petition said: “Collecting and repurposing this kind of data raises serious concerns around privacy, consent, and trust in the workplace.”
The tech publication Wired reported this week that MCI data collected from corporate laptops had been accessible to anyone inside the company. It cited an internal security notice that referred to the exposure of data tables including “full prompts and transcriptions, private conversations, people and performance data”.
Meta confirmed the program has been paused.
“We have carefully designed this program with privacy safeguards and while we have no indication at this time that any data was improperly accessed by Meta employees, we’re pausing it while we investigate,” the company said in a statement.
Zuckerberg, Meta’s founder and chief executive, had told employees that AI models – the underlying technology that powers AI tools such as Meta’s chatbots – learn from “watching really smart people do things”, according to an account of an internal company meeting.
“The average intelligence of the people who are at this company is significantly higher than the average set of people that you can get to do tasks,” he said, adding that the coding skills of Meta engineers would dramatically improve a model’s coding abilities.
Zuckerberg is pouring vast sums into an AI drive at Meta and is spending up to $145bn (£110bn) at the company in capital expenditure this year, with much of it going on AI investment such as datacentres.
Meanwhile, the New York Times has reported that Zuckerberg recently ordered a small team at the $1.4tn company to create a smartphone app similar to Polymarket and Kalshi, prediction market sites that allow users to place wagers on events ranging from Tony award winners to the Iran conflict. About $24bn in wagers in total are placed each month on Kalshi and Polymarket, according to the Pew Research Centre.
after newsletter promotion
The tentative project has been called Arena and would function separately from Meta’s social media and messaging apps, the NYT reported. It added that the proto-app remains in development and may not be released.
Mike Proulx, a research director at the analysis company Forrester, said moving into a controversial area such as prediction markets – which have already drawn legal scrutiny in the US – was “not a great look” for a company under legal pressure because of its social media products.
Meta has been contacted for comment.
Read the full story at BBC ↗ · The Guardian ↗
Meta has paused an internal program that collected employee computer activity data for AI training purposes. The program, called the Model Capability Initiative, prompted a petition from over 1,600 staff members citing privacy and consent concerns. Security researchers reported that data collected through the tool—including transcriptions and private conversations—was accessible to employees across the company. Meta confirmed the pause and stated it has found no indication of improper access but is investigating. CEO Mark Zuckerberg had previously explained the initiative by noting that Meta engineers' technical skills would significantly improve AI model capabilities. Meta is simultaneously pursuing substantial AI investments, with up to $145 billion allocated to capital expenditure this year. Additionally, the company is developing an early-stage prediction market app called Arena, though no release timeline has been set.
Read the full story at BBC ↗ · The Guardian ↗
Mark Zuckerberg’s Meta has paused a program that tracked employees’ computer activity amid data privacy concerns and a staff backlash.
The owner of Facebook, Instagram and WhatsApp had introduced a tool that tracked staff keystrokes, mouse clicks and content displayed on computer screens in order to collect data for training its AI models.
More than 1,600 Meta workers signed a petition against the tool, called the Model Capability Initiative, demanding the company does not harvest “employee ‘computer use’ data”.
The petition said: “Collecting and repurposing this kind of data raises serious concerns around privacy, consent, and trust in the workplace.”
The tech publication Wired reported this week that MCI data collected from corporate laptops had been accessible to anyone inside the company. It cited an internal security notice that referred to the exposure of data tables including “full prompts and transcriptions, private conversations, people and performance data”.
Meta confirmed the program has been paused.
“We have carefully designed this program with privacy safeguards and while we have no indication at this time that any data was improperly accessed by Meta employees, we’re pausing it while we investigate,” the company said in a statement.
Zuckerberg, Meta’s founder and chief executive, had told employees that AI models – the underlying technology that powers AI tools such as Meta’s chatbots – learn from “watching really smart people do things”, according to an account of an internal company meeting.
“The average intelligence of the people who are at this company is significantly higher than the average set of people that you can get to do tasks,” he said, adding that the coding skills of Meta engineers would dramatically improve a model’s coding abilities.
Zuckerberg is pouring vast sums into an AI drive at Meta and is spending up to $145bn (£110bn) at the company in capital expenditure this year, with much of it going on AI investment such as datacentres.
Meanwhile, the New York Times has reported that Zuckerberg recently ordered a small team at the $1.4tn company to create a smartphone app similar to Polymarket and Kalshi, prediction market sites that allow users to place wagers on events ranging from Tony award winners to the Iran conflict. About $24bn in wagers in total are placed each month on Kalshi and Polymarket, according to the Pew Research Centre.
after newsletter promotion
The tentative project has been called Arena and would function separately from Meta’s social media and messaging apps, the NYT reported. It added that the proto-app remains in development and may not be released.
Mike Proulx, a research director at the analysis company Forrester, said moving into a controversial area such as prediction markets – which have already drawn legal scrutiny in the US – was “not a great look” for a company under legal pressure because of its social media products.
Meta has been contacted for comment.
Read the full story at BBC ↗ · The Guardian ↗
Meta paused the Model Capability Initiative, a program that tracked employee keystrokes, mouse clicks, and screen content to collect data for AI model training. Over 1,600 Meta employees signed a petition against the tool, citing privacy, consent, and workplace trust concerns. Wired reported that data collected by the program—including full prompts, transcriptions, private conversations, and performance data—was accessible to anyone inside the company. Meta confirmed the program has been paused and stated no evidence of improper employee access has been found, though an investigation is ongoing. CEO Mark Zuckerberg justified the initiative by stating that Meta engineers' skills would significantly improve AI model capabilities. Meta is investing up to $145 billion in capital expenditure this year, with much directed toward AI infrastructure such as datacenters. Meta is developing an early-stage prediction market app called Arena, which remains in development and may not be released. Moving into prediction markets is 'not a great look' for a company already under legal pressure over social media products.
Read the full story at BBC ↗ · The Guardian ↗
- Meta paused its Model Capability Initiative, a program that tracked employee keystrokes, mouse clicks, and screen content to train AI models, following staff backlash and privacy concerns.
- Over 1,600 Meta employees signed a petition against the tool; Wired reported that collected data was accessible to anyone in the company, including private conversations and performance data.
- Meta confirmed the pause and said it found no evidence of improper access, but is investigating; CEO Zuckerberg had justified the tracking by saying Meta engineers' skills would improve AI models.
- Meta is investing heavily in AI (up to $145bn this year in capital expenditure); separately, the company is exploring a prediction market app called Arena, which remains in development.