AI Warning: “Social Order Could Collapse”
Japan's largest telecommunications company and the country's leading newspaper are urging swift legislation to regulate generative artificial intelligence (AI), warning of potential threats to democracy and social order if AI remains unchecked.
In a joint AI manifesto set to be released on Monday, Nippon Telegraph and Telephone (NTT) and Yomiuri Shimbun Group Holdings highlighted the need for regulatory measures similar to the law passed by the European Parliament in March, which restricts certain uses of AI. This reflects growing concerns among American allies regarding the AI technologies developed primarily by U.S.-based companies.
While acknowledging the potential productivity gains from generative AI, the Japanese companies expressed skepticism about the technology's overall impact. They cautioned that AI tools, designed to capture users' attention without ethical considerations or accuracy, have already begun to erode human dignity.
The manifesto warned of dire consequences if AI remains unregulated, including the potential collapse of democracy and social order leading to conflicts and wars.
To address these concerns, the manifesto proposed immediate actions by Japan, including the enactment of laws to safeguard elections and national security from potential abuses of generative AI.
Globally, there is a growing push for AI regulation, with the European Union taking a leading role. The EU's new law mandates safety evaluations for powerful AI models and prohibits the use of emotion-recognition AI in certain settings.
The Biden administration has also increased oversight, utilizing emergency federal powers to require major AI companies to report systems posing national security risks. Additionally, the U.S., U.K., and Japan have established government-led AI safety institutes to develop AI guidelines.
Despite these efforts, democratic governments face challenges in regulating AI-powered speech, especially on social media platforms, due to constitutional protections for free speech.
NTT and Yomiuri Shimbun emphasized that their manifesto reflects concerns about public discourse. Both companies play influential roles in shaping policy in Japan, with the government retaining a significant stake in NTT.
Yomiuri Shimbun, Japan's most widely-read newspaper, has been vocal about AI-related issues, citing risks of AI models promoting weapon-making or discriminatory content. NTT, involved in AI research, offers generative AI products and recently launched a large-language model named "tsuzumi" designed for Japanese contexts.
While acknowledging the benefits of generative AI, NTT emphasized the technology's potential risks in manipulating public opinion if misused maliciously.