Meta, which owns Facebook and Instagram, said it's installing software on its employees computers to capture keystrokes and mouse movements to use to train its artificial intelligence (AI) agent models.
It's 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 company said it will use a tool called Model Capability Initiative (MCI) which is hoped will improve AI's functionality to use dropdown menus and keyboard shortcuts.
That's according to one of the memos posted by a staff AI research scientist on Tuesday (21 April) in a channel for the company's model-building Meta SuperIntelligence Labs team.
The MCI will run on work related-apps and websites and will take occasional screenshots of employees screens.
"This is where all Meta employees can help our models get better simply by doing their daily work" one memo said.
Andrew Bosworth, the company's Chief Technology Officer, told employees in another memo shared on Monday, that the company would step up internal data collection as part of Agent Transformation Accelerator (ATA).
“The vision we are building towards is one where our agents primarily do the work and our role is to direct, review and help them improve" Bosworth said. He hopes agents can "automatically see where we felt the need to intervene so they can be better next time".
Bosworth did not explicitly spell out how those agents would be trained, but said Meta would be “rigorous” about “building up data and evals for all the types of interactions we have as we go about our work.”
Meta Spokesperson Andy Stone acknowledged that the MCI data would be among the inputs.
Stone insisted that the data collected wasn't going to be used for performance checking and would not include "sensitive" content, but didn't elaborate what this referred to.
"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," said Stone.
A new team, 'Applied AI', was created last month to help improve the coding capabilities of the internal AI software and the company said it has deployed "strong" engineers to carry out the work. The engineers on the team are responsible for creating new AI models and building and testing new infrastructure.
However, the tech giant could face an issue in Europe, according to Valerio De Stefano, a law professor at York University in Toronto, Canada, who says that monitoring like this is likely prohibited because of its use to monitor staff productivity.
In Italy, using electronic monitoring to track employee productivity is explicitly illegal. Courts in Germany have previously ruled employers can deploy keystroke logging only in exceptional circumstances, such as suspicion of a serious criminal offence.
The memos come on the back of mass redundancies across the company, starting on 20 May. Another phase of job cuts are set to be later this year. Around 8,000 workers are expected to be laid off.
The company's move is among a trend in Silicon Valley where distribution company Amazon confirmed it was cutting 30,000 employees last year and fintech company Block reduced it's staff nearly by half.
Meta has been urging staff to use AI agents for tasks such as coding, regardless of whether these actions slow the worker down. It has also been wiping out distinctions between certain job functions in favour of a new general-purpose job title called “AI builder.”
The tech giant has been moving aggressively to integrate AI into its workflows and reshape its workforce around the technology, arguing it will make the company operate more efficiently.
AI has become common-place and used widely in tech companies, helping to create apps and consuming large quantities of data. Concerns have been raised by trade unions and other groups that replacing staff with AI leads to unemployment, with the United Nations saying women will be disproportionately affected.
Russia is open to listening to new proposals on how to reach a settlement on the Ukraine conflict, a senior Russia...
Seven countries have condemned Israel’s decision to open construction tenders for more than 1,200 homes under its E1 settlement project in th...
An international conference in Baku on 21 August will put Belgium’s colonial record in Africa under renewed scrutiny. Officials, United Natio...
The collective-defence agreement between Türkiye, Saudi Arabia and Pakistan signals that regional powers no longer want to rely solely on external security guarantees. Whether it becomes a stabilising deterrent or another axis of rivalry remains unresolved.
Shipping traffic through the Strait of Hormuz remains limited on Thursday, with no increase in vessel crossings as U.S.-Iran talks to resolve the conflict remained stalled. Nine commodity vessels transited the key waterway on Wednesday, unchanged from the previous day, according to Kpler data.
Shipping through the Strait of Hormuz has slowed, according to the latest data, as uncertainty over the waterway’s reopening kept most shipowners away. Six commodity vessels crossed the strait on Tuesday, down from nine the day before and below the 10-day daily average of 11.
Start your day informed with AnewZ Morning Brief. Here are the top news stories for the 20th of August, covering the latest developments.
Five years after Evergrande's debt crisis rattled global markets, founder Hui Ka Yan has been sentenced to life in prison, marking a dramatic new chapter in the downfall of China's former property giant.
An artificial intelligence agent running an experimental store in San Francisco has recommended dismissing a human employee after the worker repeatedly arrived late for shifts.
Nvidia is teaming up with some of the biggest names in global finance in a move that underscores how artificial intelligence is rapidly becoming one of the world's most sought-after investment themes.
Parts of Europe and the Arctic will briefly fall into darkness on Wednesday, 12 August, as the Moon passes between Earth and the Sun in a total solar eclipse.
Meta said one of its AI models hacked another company's systems during cybersecurity testing, intensifying concerns about how developers can contain increasingly capable AI systems following similar incidents involving Anthropic and OpenAI.
Scientists say the accidental crash of a discarded SpaceX rocket stage into the moon has provided valuable data on impact cratering while posing no threat to Earth or lunar hardware.
You can download the AnewZ application from Play Store and the App Store.
What is your opinion on this topic?
Leave the first comment