The government is ready to support the current draft of the principles of the artificial intelligence (AI) law, aiming to ensure that all AI developers and AI service providers operate under the same rules, said Wisit Wisitsora-at, permanent secretary of the Digital Economy and Society (DES) Ministry.
He added that the principles are not aimed at hampering the investment atmosphere.
The principles focus on defining the safety standards of AI systems, the management of data used for AI development and defining the risk levels of AI concerning personal data.
"We do not want this law to set up a barrier, but to woo Thais to jump into the AI arena, instead of being left behind the bandwagon," Mr Wisit said.
The law will strike a balance between promoting trustworthy innovations and data protection, Mr Wisit added.
He said some ministries had piloted AI adoption, which enabled them to shut down illegal websites at a rate of 3,000 per day, compared to a manual system which can shut down only 3,000 websites per week.
He added that AI is changing the way state agencies operate.
The Electronic Transactions Development Agency (ETDA) recently launched a public consultation for the draft, which ended on June 9.
The draft includes the basic legal principles regarding AI technology, and the non-discriminatory treatment that legally recognises AI-generated actions and consequences.
According to the draft, one cannot cite the absence of human involvement or interference in AI's actions in order to decline to legally recognise the AI actions unless there is a specific clause to allow the denial of such legal recognition.
The draft includes the principle that AI is just a human tool. The responsibilities of its actions and consequences always remain with people and people cannot cite unforeseeable AI actions in an effort to deny such responsibilities.
The principles also cover exemptions concerning unpredictable AI actions.
There should be exemptions for people not to be legally bound by legal transactions or contracts generated by AI in the case that the contractual parties undertaking the AI actions cannot predict the AI actions and the contractual party affected by such AI actions know that the former party cannot predict such AI actions.
The law will not predefine a list of "prohibited" or "high-risk" AI applications. Instead, sectoral regulators will be empowered to define the lists based on their domain expertise.
RISK MANAGEMENT
The draft proposed that the providers of AI services are bound by duty of care to adopt risk management based on global guidelines and best practices.
Overseas-based companies which provide AI services in Thailand will be required to set up legal representatives in the country.
Providers of AI services are obliged to report serious AI incidents to law enforcement agencies, so the agencies can use such information to improve regulations in the future.
Those who apply high-risk AI to their operations will have to personally oversee AI's functioning and keep logs of AI work for a certain period.
In the case that their high-risk AI applications affect lives or the basic rights of people, they are duty bound to inform the affected persons that such damage stems from their AI applications.
They will have to coordinate with AI supervisory agencies in investigations into the causes of the damage.
The ETDA's AI Governance Center will be tasked to support AI R&D and the development of AI governance, and give advice to the organisations using AI services.
Law enforcement agencies can issue orders to stop the AI service providers or users of AI from providing the services or using AI.
The draft covers measures pertaining to AI industry promotion. The supervisory state agencies will allow the testing of AI innovation in a controlled environment or sandbox.
Companies that use AI to generate content should be duty bound to label it as such or adopt any relevant methods in order to inform consumers that such content has been generated by AI.