Meta planning to launch agentic AI assistant for 3 billion users, report says

Meta is reportedly developing personalised AI agents for more than 3 billion of its users. These agents may work similar to OpenClaw, where you can automate your tasks. Do note that this comes at a time when Meta is planning mass layoffs to save funds for investments, likely in AI.

In Short

• Meta is reportedly building new personalised AI agents
• The Meta agents likely to be similar to OpenClaw
• Meta may allow users to share sensitive data to these agents

AI agents are growing in popularity. You may have heard about OpenClaw and how AI agents can automate tasks for you. For the most part, AI agents are used for writing code or doing work. But it seems that Mark Zuckerberg wants you to use AI agents even for your everyday tasks.

As per a report from the Financial Times, Meta is working on new personalised AI agents that will be able to do everyday tasks for more than 3 billion of its users.

Do note that this comes at a time when Meta is investing billions of dollars in AI. The company has also announced layoffs that will impact roughly 10 per cent of its global workforce in a bid to fund its new investments.

A Meta rival to OpenClaw

OpenClaw became quite popular earlier this year, particularly after the launch of Moltbook – an AI-only social media platform where OpenClaw bots were demanding a takeover of the world. Meta recently acquired Moltbook, with the Moltbook team joining the Meta Superintelligence Labs. As per the report, Meta’s new personalised AI agents will likely be similar to OpenClaw.

Unlike OpenClaw, which can be powered by different AI models, depending on the user, Meta’s agents are expected to be powered by its own Muse Spark AI model – the first model to be developed by its Superintelligence Labs led by Alexandr Wang.

What will Meta’s AI agents do?

It is believed that Meta’s new agents are being tested internally by a group of staff. Similar to OpenClaw, users will likely be able to assign tasks to their agents, who can then do the work in the background. It is unclear, however, if the agents will run locally on the device, akin to OpenClaw, or whether everything will happen via the cloud, more like Anthropic’s Claude.

The report indicates that Meta wants users to be able to share sensitive information with these models, including health and financial data. Do note that last year, Meta settled an $8 billion lawsuit over alleged user privacy violations on Facebook.

AI Mark Zuckerberg, employee tracking and more

Meta’s push for AI is not limited to AI agents. The company has been doubling down on AI for some time now. Apart from AI agents, Meta is rumoured to be planning to integrate a separate agentic shopping tool into Instagram and is targeting a launch before the fourth quarter of this year.

On the organisational side, there is AI too. It is believed that Mark Zuckerberg now has his own AI avatar, which employees can interact with. Recently, Meta also announced that it was going to track employee mouse movements and keystrokes to train AI. With the aim of making AI better at work in the long term.

Last week, Meta stated that it was raising its capital expenditure by $10 billion to $145 billion this year with a big focus on AI. However, this was met with caution by Wall Street. Meta’s market cap fell by almost $170 billion following the announcement.

The Mark Zuckerberg-led firm also attempted to acquire Chinese startup Manus AI for $2 billion. Though the deal was blocked by Beijing.

- Ends