Discuss Deepfakes:
The director of the institute said the institution could help the military with cognitive warfare countermeasures.
-
Shelley Shan / Reporter
The National Science and Technology Council (NSTC) on Monday proposed a draft artificial intelligence (AI) basic law aimed at managing the development of AI applications and the associated risks.
The draft bill stipulates that governments should classify AI applications based on the risks associated with their use and promote trust in them through standards, validation procedures, and other accountability mechanisms.
It also sets out the conditions under which AI technology can be applied, as well as issues relating to liability, remedies, compensation and insurance.
Photo: Chen Yi-kuan, Taipei Times
According to the draft, AI applications will not be subject to regulation during the research and development stage.
The agency responsible for protecting personal data says it should help all government agencies avoid unnecessary collection, handling and use of personal data during the research and development stage of AI applications.
The bill states that the government should prevent the use of AI applications that lead to loss of life, liberty or property, disruption of social order, damage to the environment, conflict of interest, discrimination, false advertising, misleading or fabricated information.
Under the draft bill, the Digital Ministry will be required to recommend tools and methods that government agencies can use to assess and verify the risks associated with the use of AI applications.
Government agencies will be responsible for managing the use of AI applications, including requiring foreign AI service providers to set up offices in Taiwan.
The ministry should also develop an AI risk management framework similar to that developed by the U.S. National Institute of Standards and Technology, which government agencies can use to build their own risk management mechanisms for AI applications, the draft said.
The government also said it should help workers who lose their jobs to AI find new jobs.
The government will be empowered to provide subsidies, funds, tax incentives and other measures to promote research and application of AI technologies and the construction of infrastructure.
The bill was enacted after extensive discussions with industry leaders, university deans and government officials, and is currently open for public comment until Sept. 13, the council said.
In other news, the Zhongshan Institute of Science and Technology announced yesterday that it now has research and development capabilities to help the military counter cognitive warfare, which has intensified in recent years due to deepfake technology.
“Deepfake technology – a term that combines “deep learning” technology and “fake” – is a challenge and risk arising from the widespread application of AI technology,” said Huang Yi-jin, director of the institute's information and communications research division.
“Deepfake technology can be used to create lifelike photos, audiovisual content and text that can be quickly spread over the internet. It can also be used in cognitive warfare between hostile countries,” Huang told a news conference at Taipei's Ministry of Transportation and Communications.
In cognitive warfare, the targeted audience may believe what is fabricated to be true, leading to internal conflict and low morale, he said.
He said the technique, when combined with phishing messages or emails, could also be used to fraudulently obtain personal or confidential information of specific individuals.
Asked whether the institute could use deepfake technology to help the military counter China's cognitive warfare against Taiwan, Huang said the institute had the research and development capabilities to assist the military in this regard.
He did not elaborate.
Defense Ministry spokesman Major General Sun Lifang said much of the misinformation was circulating during the Hankuan military exercises, saying the exercises were designed to help senior military officers escape in the event of war or that serious mistakes had occurred, resulting in casualties.
Sun said the joint efforts of government agencies have enabled people to better distinguish fake news from the truth.
“We intend to control the narrative surrounding the issue and stop it before it spreads,” he said.
Comments are moderated. Please keep your comments relevant to the article. Comments containing insults, obscene language, personal attacks of any kind or promotion will be deleted and users will be banned. The final decision is at the discretion of Taipei Times.
