Calls for global regulation of AI are growing. What exactly could it look like?

Olga Tsyvinska/Getty Images Australian Prime minister Anthony Albanese appears to be listening to artificial intelligence (AI) firms’ warnings to the world about the dangers of the technology they are developing.

Over the weekend, he said global rules were needed to regulate the development of AI “to make sure that humans remain in control”.

This comes after Dario Amodei, the boss of AI firm Anthropic, called for AI development to slow down so governments and firms have more time to address the “serious” risks associated with the technology.

Other leading AI figures, such as Sam Altman, the head of OpenAI, agree with him .

But how exactly could AI be better regulated? And who exactly would enforce it? What’s prompting this? There are growing concerns about the ability of AI to escape human control.

This comes after a series of high-profile AI cybersecurity incidents.

For example, just last week Google announced its Gemini AI model autonomously hacked three websites during a test of its cybersecurity capabilities in May.

Similar incidents involving Anthropic’s and OpenAI’s AI models have also occurred recently.

One way to interpret these cybersecurity incidents is as hyperbolic marketing.

In a world of rapid media cycles, they help maintain the excitement on AI capabilities.

They are also far more interesting than incremental tech development.

But they also raise the question of whether AI firms need greater oversight and regulation.

What are countries currently doing? Last week, the Australian government released a consultation paper on proposed national “AI standards”.

The focus is on realising opportunities and minimising potential harms around data centres and AI training infrastructure.