Japanese telecom leader NTT and national newspaper The Yomiuri Shimbun have issued a joint proposal warning that unchecked development and deployment of generative AI could inflict “enormous and irreversible damage” on society. While recognizing the technology’s practical benefits, the two organisations call for urgent governance measures to manage the risks that accompany rapid adoption.
The proposal acknowledges generative AI’s clear advantages: intuitive user interfaces, faster workflows, and productivity gains from summarisation and information retrieval. At the same time, it highlights serious concerns about the diminishing role of human oversight and the growing difficulty of ensuring the accuracy, fairness and safety of AI-generated content.
Key risks identified include model hallucinations that produce false or misleading statements, systemic bias and toxic outputs, and copyright infringement arising from large-scale data scraping. The paper also warns that if AI erodes incentives to produce and verify accurate information, the broader information ecosystem — and public trust in it — could be severely damaged.
“If generative AI is allowed to go unchecked, trust in society as a whole may be damaged as people grow distrustful of one another and incentives are lost for guaranteeing authenticity and trustworthiness,” the proposal states. In their strongest terms, the authors caution that unchecked harms could, in a worst-case scenario, threaten democratic institutions and social order.
To reduce these dangers, NTT and Yomiuri advocate a balanced approach that combines legal safeguards, technical controls and industry self-regulation. They propose restoring meaningful human autonomy and dignity in interactions with AI, restricting or tightly governing AI use in sensitive areas such as elections and national security, and updating intellectual property frameworks to reflect the realities of AI training and generation.
The proposal recommends layered governance: enforceable laws or strict regulation for high-risk applications alongside lighter, industry-led co-regulation for less critical domains. This hybrid model aims to provide clarity and accountability where failure would have grave consequences while allowing innovation to continue in lower-risk contexts.
NTT and Yomiuri stress that failure to act could permit the “unconditional application” of generative AI in domains where errors or misuse have outsized consequences, potentially producing social effects that are difficult or impossible to reverse. They call for immediate steps to establish clear rules, technologies and norms that preserve democratic processes and public safety.
The paper urges a multi-pronged response: legislation to set baseline obligations and prohibitions, organisational and industry rules to operationalise responsibility, collaborative mechanisms for cross-sector governance, and technical safeguards such as transparency, robust auditing, human-in-the-loop designs and provenance tools to trace and verify AI outputs.
Recognising generative AI’s potential to drive innovation, the proposal nevertheless frames governance not as a brake on progress but as a necessary condition for sustainable adoption. By creating a “healthy space for discussion” and embedding protections across technical, legal and social layers, policymakers and industry can help ensure AI contributes positively while minimising harm.
The initiative is supported by Keio University’s Cyber Civilization Research Center, which is collaborating with NTT and Yomiuri to study the technology’s societal impacts and to develop concrete policy recommendations for responsible development and governance.
Read the full joint proposal (PDF) from NTT and The Yomiuri Shimbun for detailed recommendations and analysis.
(Photo by Steve Johnson)
See also: Virgin Media O2 launches ‘smart support’ to proactively solve issues
Interested in learning more about AI and big data from industry leaders? The AI & Big Data Expo runs events in Amsterdam, California and London. These conferences bring together practitioners across AI, cloud, cybersecurity and digital transformation to share strategies, case studies and best practices for responsible technology deployment.
Explore additional enterprise technology events and webinars from TechForge to stay informed about emerging trends, governance approaches and technical safeguards that help organisations use AI safely and effectively.