Microsoft Thailand, the local operating unit of the global technology firm, is urging policymakers and regulators to get ready for the dark side of artificial intelligence by endorsing guidelines and principles to protect users and ensure that AI is trustworthy.
An ongoing discussion with the company, the Bank of Thailand and the Stock Exchange of Thailand is meant to raise awareness of the issue.
"AI adoption is ready to take off in Thailand, and it's important to note that the government needs to prepare the regulatory environment to ensure innovation development and user protection," said Ome Sivadith, national technology chief of Microsoft Thailand.
There have been cases of AI misuse, such as voices generated by AI to mimic others' voices, including a fake video of former US president Barack Obama.
AI algorithms can discriminate based on skin tone and region, depending on how the machine was trained to learn and input data.
A fitness tracker that traces the user's running can also generate map locations from those footsteps, or provide mapping in a house where smart home appliances are used, Mr Ome said.
He said policymakers need to have dialogue with multiple stakeholders to declare AI principles or best practices, fund R&D and share the results.
Principles for trustworthy and human-centred AI need to be practical, he said.
Policymakers also need to consider accountability for developers in case there is an AI malfunction, such as an autonomous vehicle crashing into a person. Legal standards for responsibility and accountability need to be updated.
Mr Ome said banks historically have used credit scoring systems with humans making decisions. If the machine is the sole assessment method, AI should be transparent and able to explain where the result comes from.
"The Monetary Authority of Singapore is working on guidelines for AI to be used by financial institutions, especially fintech services," Mr Ome said.
Policymakers need to train new workers about AI and re-skill existing employees, he said.
