Musk’s Grok AI faces scrutiny from Japan over sexualised images

TOKYO – Japan joined the growing list of countries probing X over tech billionaire Elon Musk’s artificial intelligence service Grok and the chatbot’s role in creating and spreading sexualised images of people without their consent.

The Cabinet Office has asked the social media platform to improve safeguards and curb the output of sexually altered images by Grok, Japan’s Economic Security Minister Kimi Onoda said.

Officials also submitted written questions about measures by X to prevent the creation of deepfakes and other images that violate people’s privacy, intellectual property and right to control the use of their likeness, she said.

Grok in January drew a firestorm of

rebuke from users and governments from Malaysia to Italy

over the ease with which it could be used to manipulate photos to sexualise or demean their subjects. It has since applied some restrictions on its image-generation capabilities and put them behind a paywall.

“All possibilities, including legal steps, are on the table if there is no improvement,” said Ms Onoda, who is also minister of state for AI strategy, at a regular news conference on Jan 16.

“If similar problems occur on other platforms, we will also act.”

The statements, coming from a minister in X’s biggest overseas market, follow investigations from regulators around the world. Canada, California, the EU and individual member states like France are probing whether Grok’s generated images violate people’s rights, while Malaysia, Indonesia and the Philippines have restricted domestic access.

Earlier this week, xAI – the company responsible for Grok –

sa

id it is

disabling the ability for people to use the Grok chatbot

to create sexualised images of real people.

“It’s a problem that people can still create such images,” Ms Onoda said, suggesting Japan may seek further measures from the chatbot creator. But she also added that the fault is not in AI itself, and that the government needs to study different ways to address the real issues.

“If a knife is lying there, do you use it to cook or to hurt someone?” she added.

Japan is balancing a need for guardrails on AI with its ambition to catch up to the US and China in an area it sees as essential for national strategy. Japan’s AI law, which took full effect in September 2025, carries no penalties, limiting government interference to investigations and issuing formal guidance when violations occur.

Discussions among Japanese policymakers have centred around ideas such as ways to educate users and make AI services show that images have been doctored, although the Cabinet Office is also seeking to coordinate possible measures with the country’s Justice Ministry, the National Police Agency and the Communications Ministry. BLOOMBERG