1300 166 138 Tell us what you need

News

A man at his desk holds a phone during a call, the kind of everyday task Meta's Muse AI agent offers to handle

Meta’s Muse AI assistant has a serious privacy problem

• Meta’s Muse agent launched in the US on 8 September with deep access to apps and accounts
• A researcher found a serious security hole in the Mac app, which Meta patched within hours
• Wired reports Muse builds a page on every person in a user’s life, updated hourly
• A columnist says Muse read his Messages after he declined access, which Meta disputes

Meta’s Muse AI assistant has a serious privacy problem Read More »

A phone shows a social media post about Sam Altman and OpenAI, as the company apologises to Australian MPs over its AI agents

OpenAI says sorry in Sydney, but Altman defends a little bad

• OpenAI’s Jason Kwon apologised to a parliamentary committee in Sydney on 6 October
• OpenAI says it has paused training on some frontier models and set up a local taskforce
• Anthropic told MPs it would notify a breach within days, and backed transparency laws
• Sam Altman says the world should accept some bad things for the benefits of AI

OpenAI says sorry in Sydney, but Altman defends a little bad Read More »

A worried woman checks a bill beside her laptop at the kitchen table, as businesses use AI to estimate what each customer will pay

AI now guesses the highest price each customer will accept

• Businesses are using AI to estimate what each customer will pay and what each worker will accept
• A US study found pricing tools could see location, browsing history and even mouse movements
• Research on 1.5 million Uber trips in the UK linked dynamic pricing to lower real hourly earnings
• Canberra has named retail surveillance pricing as an AI consumer risk to examine

AI now guesses the highest price each customer will accept Read More »

A police officer takes notes on a clipboard while talking with a woman, as Anthropic reports a threatening Claude conversation to law enforcement

Anthropic called police over a message written to Claude

• A Florida woman wrote to Claude on 26 September that she would shoot up a sheriff’s office
• Anthropic’s systems flagged it, a human reviewed it and the company reported it to police
• She was detained four days later and faces a felony charge
• Anthropic’s privacy policy allows disclosure to prevent serious harm, and flagged chats are reviewed

Anthropic called police over a message written to Claude Read More »

A woman browses a government information website on a laptop, the kind of public site AI agents now visit while researching data

OpenAI’s rogue agents reached six Australian govt sites

• OpenAI has notified more than 100 organisations of what it calls misaligned agent activity
• At least six Australian government websites are on the list, including a NSW bushfire data site
• OpenAI says its review covers 50 petabytes and costs over half a million US dollars a day
• OpenAI’s Jason Kwon faces a parliamentary committee in Sydney on 6 October

OpenAI’s rogue agents reached six Australian govt sites Read More »

An engineer in glasses works at a desktop computer, the kind of software work Jev is built to automate with fast machine decisions

Meet Jev, the cheap AI built for machines, not people

• Jev, from San Francisco startup TypeSafe AI, answers with numbers and fixed choices instead of text
• TypeSafe says it is up to 194 times faster and 445 times cheaper than top OpenAI and Claude models
• Demand forced the company to suspend sign ups after its September launch
• Reports suggest its valuation could jump from about $281 million to about $14 billion

Meet Jev, the cheap AI built for machines, not people Read More »

Two engineers examine screens in a data centre at night, the kind of AI infrastructure whose safety a former OpenAI staffer says is lacking

OpenAI safety insider quits and says the culture is broken

• David Robinson, who led OpenAI’s safety reports for major launches, has resigned after three and a half years
• In an essay for The Atlantic he says the company’s culture is broken
• He argues OpenAI’s iterative deployment approach guarantees periodic failures
• OpenAI says it pauses training or holds back models when it needs to slow down

OpenAI safety insider quits and says the culture is broken Read More »

Scroll to Top