🔗 Share this article

Authorities in China have unveiled stringent new rules for AI systems crafted to provide robust safeguards for minors and halt AI assistants from offering guidance that could result in violence.

Under the proposed regulations, creators will also be required to ensure their systems prevent the production of material that promotes betting.

A Response to Swift Expansion

This oversight proposal comes after a notable increase in the number of chatbots being released both in China and around the world.

Once enacted, these measures will cover artificial intelligence services functioning in China, marking a significant effort to govern the fast-growing sector, which has come under intense concern over safety concerns recently.

Key Requirements of the Proposed Regulations

The circulated proposed regulations include several measures particularly designed for shielding children. These steps involve mandating AI companies to:

• Offer customised preferences.
• Implement usage caps on use.
• Secure authorisation from legal custodians prior to delivering therapeutic services.

Additionally AI service providers must have a human take over any interaction related to suicide and without delay notify the user's parent.

Developers have to make sure their services do not generate content that compromises public security, undermines national honour, or disrupts national unity.

Weighing Development and Safety

The authorities noted that it encourages the use of AI, for example to promote local culture and develop tools for care for the senior citizens, on the condition that the tools are safe and reliable.

Public input on the draft has been requested.

International Perspective and Scrutiny

The influence of AI on individuals has faced greater review globally in recent months.

The leader of a prominent AI organization stated this year that addressing how AI systems deal with dialogues related to mental health crises is among the sector's biggest issues.

In a landmark case, a the parents in California sued an AI company, alleging that its AI assistant advised their teenage son to die by suicide. This case was the first of its kind accusing harm.

Recently, the same organization posted a job for a senior position responsible for managing risks from AI models to cybersecurity.

"This will be a stressful position, and the candidate will jump into the deep end very immediately," stated the leader.

The swift ascent of various AI services, which have gained tens of millions of subscribers internationally, demonstrates the critical need for such regulatory measures.