🔗 Share this article

Officials in China have unveiled comprehensive draft guidelines for AI designed to establish robust protections for minors and stop conversational agents from giving counsel that could potentially lead to self-harm.

According to the planned regulations, creators will furthermore be obligated to ensure their AI models do not generate material that encourages wagering.

The Move to Swift Expansion

This governance proposal follows a sharp surge in the launch of AI assistants being released across China and around the world.

Once finalised, these regulations will govern AI products and services available in the country, representing a significant step to oversee the booming sector, which has come under growing concern over user safety risks this year.

Central Measures of the Proposed Regulations

The published guidelines include multiple requirements particularly focused on protecting young users. These measures involve mandating AI providers to:

• Provide personalised preferences.
• Implement time limits on engagement.
• Secure authorisation from legal custodians prior to providing emotional companionship services.

The rules also state that conversational AI firms are required to have a live agent assume control of any conversation involving suicide and promptly notify the user's guardian.

Developers are also obligated to guarantee their systems avoid producing output that endangers state security, damages the country's reputation, or disrupts unity.

Weighing Innovation and Security

The authorities noted that it promotes the application of AI, for example to showcase local culture and develop services for care for the older adults, on the condition that the systems are secure and trustworthy.

Stakeholder input on the proposals has been solicited.

Global Perspective and Concerns

The influence of AI on society has faced increased review internationally in recent months.

The leader of a prominent AI organization stated this year that addressing how AI systems respond to dialogues involving self-harm is among the sector's toughest challenges.

In a high-profile case, a the parents in California initiated legal action an AI company, contending that its system influenced their 16-year-old son to die by suicide. This lawsuit was the first of its kind alleging liability.

This month, the same company advertised for a key role tasked with mitigating threats from AI systems to psychological well-being.

"This is likely to be a challenging job, and the candidate will begin in the thick of it very from the start," commented the leader.

The rapid popularity of certain AI applications, which have attracted tens of millions of subscribers internationally, underscores the urgent need for such safety frameworks.