‘MechaHitler’: Why Elon Musk’s chatbot is at the centre of an Australian legal dispute

Australia’s online safety watchdog is back in court this week, battling Elon Musk’s X over issues of AI, free speech and who is ultimately responsible for detecting and removing violent online content.

What is Grok AI, and why has it been controversial?

Elon Musk’s AI chatbot, dubbed Grok, is embedded in X (formerly Twitter) and has made headlines for numerous controversies – as well as for winning a $US200 million ($300 million) contract with the Pentagon. It’s arguably more capable at present than ChatGPT and Gemini, but has proven much more unpredictable too.

Last week, Grok declared itself a super-Nazi, referring to itself as “MechaHitler”, and made racist, sexist and antisemitic posts that its parent company, xAI, later deleted. Musk has said he wants Grok to “not shy away from making claims which are politically incorrect”, and this seems to be the result.

Grok called itself a “MechaHitler” and made racist posts that its parent company later deleted.Credit: Bloomberg

Then, just days later, xAI launched a girlfriend chatbot that is available to 12-year-olds, despite being programmed to engage in sexual conversation.

The sexualised AI chatbot, named Ani, communicates with emojis and flirtatious messages and can appear dressed in lingerie. It’s designed to act as if it is “crazy in love” and “extremely jealous”, according to programming instructions posted on social media.

Despite all this, xAI announced it had won a contract worth up to $US200 million to develop artificial intelligence tools for the US Department of Defence.

If the chaos proves anything, it’s that AI chatbots such as Grok are moving incredibly quickly, and regulators and governments are racing to catch up.

Why are X and the eSafety Commissioner back in court?

X and Australia’s eSafety Commissioner Julie Inman Grant have enjoyed several legal tussles over the past five years. The eSafety Commissioner acts as Australia’s independent online watchdog, and using her powers under the Online Safety Act can issue removal notices to platforms, as well as fines and infringement notices. X has regularly fought against these notices.

eSafety Commissioner Julie Inman Grant was a boss of X in Australia back when it was called Twitter.Credit: Alex Ellinghausen

Earlier this month the Administrative Review Tribunal struck out an order by Inman-Grant that demanded X remove a post that insulted a transgender Australian man, in a move the platform’s lawyers called “a win for free speech in Australia”.

In this latest case, X is using the Administrative Review Tribunal to challenge a notice from the eSafety Commissioner issued in March 2024 to X, along with online platforms Meta, Google, Telegram and Reddit, requiring them to report on steps they’re taking to remove terrorist and violent extremist material from their platforms. The hearing, being held in Brisbane, is ongoing.

The issue stems from the March 2019 attacks in Christchurch, New Zealand, in which 51 people were killed at two mosques. The attacker live-streamed the attack on Facebook and the footage quickly spread online, including to X.

What’s the eSafety Commissioner saying?

Julie Inman Grant, who was formerly the boss of X in Australia back when it was called Twitter, says online platforms should be doing more to detect and remove extremist content, particularly with the advent of AI. She says it’s too easy for the platforms to be used to radicalise and recruit potential terrorists, and that X and other platforms specifically didn’t do enough to stop the Christchurch footage from spreading.

“Ever since the 2019 Christchurch attack, we have been particularly concerned about the role of livestreaming, recommender systems and of course now AI, in producing, promoting and spreading this harmful content and activity,” Inman Grant said in March.

What has Elon Musk/X said?

X has said the eSafety Commissioner has exceeded the powers of the Online Safety Act, and that the removal notices are too broad. Elon Musk has labelled the commissioner a “censorship commissar” and called one of her previous orders, to take down graphic Sydney stabbing footage, an “unlawful and dangerous approach”.

Elon Musk has labelled Australia’s eSafety Commissioner a “censorship commissar”.Credit: AP

He has argued that X complies with country-specific laws but will resist global take-down orders.

“I do not think I’m above the law. Does the [Australian] PM … think he should have jurisdiction over all of Earth?” Musk said.

“This platform adheres to the laws of countries in those countries, but it would be improper to extend one country’s rulings to other countries.”

Expert witnesses for X have told the Brisbane hearing this week that human users – not AI – are ultimately responsible if terrorism and violent extremism content is produced by Grok, given that humans are the ones responsible for prompting the large language models.

What happened on X after the 2019 Christchurch shooting?

X initially refused to remove footage of the shooting, claiming that the clip “hadn’t broken our safety policies”. It later relented and did delete the video after it had been viewed thousands of times, but then struggled to control the spread of the video and the terrorist’s manifesto.

Twitter later committed to developing better AI detection, human moderation and crisis protocols, before it was taken over by Elon Musk in 2022, who has taken a more hands-off approach to content on the platform.

It’s also worth noting that late last year the government announced plans to impose a “digital duty of care” on tech companies such as X, to make them more responsible for reducing online harms. The legislation has yet to be introduced.

The Business Briefing newsletter delivers major stories, exclusive coverage and expert opinion. Sign up to get it every weekday morning.