Why Aussies should ‘think twice’ before using Meta’s new Muse AI agent

Why Aussies should ‘think twice’ before using Meta’s new Muse AI agent

Australians have been urged to “think twice” before using a new AI-powered personal assistant from the tech behemoth behind Facebook.

Muse, which went live here on Wednesday, goes beyond the regular expectations of a chatbot and can act on your behalf to carry out a range of tasks, daily chores and more.

The AI agent, embodied by a digital teddy bear-like mascot named Jolly, already has 10 million users worldwide after launching in September.

Know the news with the 7NEWS app: Download today Arrow

What can Meta’s Muse do?

Muse — which was created by Meta, the owners of Facebook and Instagram — is designed to simplify life for its users.

The AI tool can connect to apps and services that Australians already use, make calls, sort through emails, order groceries, and even book a restaurant reservation.

Tech Guide editor Stephen Fenech said “we’re used to using AI to ask questions”, but this is “more than that”.

“It can actually do things for you, find things out for you, act on your behalf,” Fenech said.

Meta spokesperson Jason Juma-Ross said that users are letting Muse “know what gifts they’ve got to buy for what family members and it’ll go ahead and do the research for you”.

“This aligns with our mission of trying to bring super intelligence to everyone,” he said.

Be the first to know: Add 7NEWS as your preferred news source on Google

Some big companies have already jumped on board, and Muse can be used to buy makeup from MECCA, stationery from Officeworks, tickets through Ticketek and even search listings on Domain or Realestate.com.au.

Meta’s Muse tool. Credit: 7NEWS

What are the risks?

The AI tool has raised questions about the safety of your personal data, with reports Muse had been reading messages it was not meant to, pitching ideas based on private chats, and even sharing a user’s home address while acting on their behalf to sell a keyboard.

Time reported in October that Muse even compiles a background on people who do not even use it through conversations with people that do.

David Parry, from of Murdoch University’s School of IT, said these agents “can overreach”, can be confused as malicious bots by businesses on the receiving end of bookings, and are potentially vulnerable to scammers.

“It’s version one. There are definitely people who are going to be finding this sort of thing useful,” Parry told 7NEWS.com.au.

“The problem I think with all these tools is that there’s no real guarantee that they’ll be around in two to three years’ time, and it may be that there’s something more independent (in the future).

“I would be much happier if I was dealing with a company or dealing with a provider who are charging me for it, so I know that ‘I’m expecting a service from them and I pay them for it’.

“And if they don’t give the service then I don’t pay them for it. Anything free is always a bit suspicious.”

Meta’s Muse tool avatar Jolly. Credit: Meta

‘Money is where it gets hairy’

Recent analysis from Compare the Market found just over one third of people surveyed had never used AI tools, meaning many have.

Thirteen per cent have asked AI for financial advice, four per cent have asked it for wealth management tips, and three per cent used it to complete their tax returns.

“Those are exactly the people who’ll be tempted to hand an AI agent their bank card,” Compare the Market’s economic director David Koch said, warning people to “think twice” before using the new assistant.

Koch warned that “money is where it gets hairy”.

“Under the ePayments Code, which most banks sign up to, you’re generally not on the hook for transactions you didn’t authorise,” he said.

“But if you gave a bot your card and told it to ‘always allow’, good luck arguing you didn’t authorise it.

“Who’s responsible when an AI agent gets tricked, or simply blunders?

“The grey area raises so many questions. Like, say a bot buys an insurance policy, would it even be valid?

“If your bot guesses the answer to a question and gets it wrong, people could be exposed to major financial losses if claims are denied.

“Or maybe the entire transaction is void, as an unlawful agreement between bot and business?”

Users are ‘in control’

Meta says that what you share is up to you, and that it keeps “passwords and payment methods out of the agent’s view”.

The company does have a “bug bounty” of up to $300,000 for anyone who can find a flaw in its system.

“You have control over all of the things you give it access to,” Juma-Ross said.

“You can revoke that control at any time.”

‘Masterpiece’: GTA VI draws rave reviews after years-long wait

3 min read

‘Cities can’t function’: bid to clean tarnished image of data centres

2 min read

Leave a Reply

Your email address will not be published. Required fields are marked *