The Allure of AI Must not Make Us Forget Its Risks

Various concerns surround generative artificial intelligence, which makes texts and images that appear to have been created by humans. If the government promotes its use too easily, it may sow the seeds of future of trouble.

The government’s AI Strategic Council, chaired by Prof. Yutaka Matsuo of the University of Tokyo, has compiled points of discussion for AI-related policy issues.

As the changes brought about by AI appear likely to surpass those of the Industrial Revolution and the internet revolution, the council has characterized AI-related changes as “the arrival of a great opportunity” for Japan.

But the council also cited seven risks of generative AI, including inappropriate use of personal information, a flood of false information, increasingly sophisticated crimes and copyright infringement. It stressed that the utilization of AI and the response to the risks should be “facilitated in a well-balanced manner.”

However, the impact that generative AI will have on society is extremely serious.

The council said in its points of discussion that “existing legal systems should be made well known to the public,” but the risks are not so slight that such a measure alone can balance the utilization of AI and the response to its risks. Priority must be placed on newly formulating regulations and legislation.

There are growing voices around the world claiming copyright infringement by generative AI. A group of artists and illustrators in the United States has filed a class action lawsuit against AI operating companies, claiming that their works were imitated by AI without their consent.

Robert Thomson, chief executive officer of the company that publishes The Wall Street Journal, warns that the rise of generative AI could “fatally undermine journalism.”

Japan revised its Copyright Law in 2018 to allow AI to freely “learn” texts, images and other copyrighted works without the consent of rights holders. Isn’t it a serious mistake to have eased restrictions on AI in contrast to other countries? The government should consider revising the law again.

Another important issue is how to prevent inappropriate use of personal information. As generative AI, which requires the input of entire texts, is more likely to reveal personal interests and tastes than internet search engines that operate based on separate words, there is no escaping the fear that personal information could be passed on to AI service providers.

The use of AI in the medical field also must be advanced cautiously. It is not clear who will be held responsible if, for example, an AI misdiagnoses a patient’s illness.

The points of discussion were drawn up barely more than two weeks after the inauguration of the AI Strategic Council — which only met twice. It is also problematic that Prime Minister Fumio Kishida has instructed that the points of discussion be reflected in the outline of the budget proposal for next fiscal year, which will be compiled in mid-June.

Based on the harsh experience of Japan having fallen behind other countries in terms of digitization, the government wants to restore the nation’s position with regard to AI. But it is wrong to rush headlong into utilizing it while leaving consideration of the risks on the back burner.

(From The Yomiuri Shimbun, June 2, 2023)