Negotiations Underway for Advanced AI Tools
The Japanese government is currently engaged in discussions with the U.S.-based artificial intelligence firm, Anthropic, aiming to secure access to its newest AI model. This push comes amid increasing international apprehension regarding sophisticated cyberattacks that utilize advanced AI capabilities.
Focus on the Powerful ‘Mythos’ Model
The urgency stems largely from Anthropic’s Claude Mythos model. This system has reportedly demonstrated a high proficiency in identifying vulnerabilities within various software programs and web browsers, leading to fears that it could be exploited for attacks across numerous different systems.
National Security Objectives
According to government sources, Japan intends to utilize this advanced AI technology to significantly enhance its defenses against cyber threats. Specifically, the goal includes mitigating risks from hostile state actors, including those believed to originate from China and Russia. Furthermore, securing access would allow Japan to gain valuable knowledge and expertise in the rapidly developing field of AI development itself.
Government Directives and Countermeasure Planning
In response to these escalating digital threats, Prime Minister Sanae Takaichi directed Hisashi Matsumoto, the minister in charge of cybersecurity, on the same day. The instruction was for him to develop and consider comprehensive measures to counter cyberattacks facilitated by the latest AI tools.
During a press conference following a Cabinet meeting, Matsumoto confirmed that Japan is “exchanging views with the United States” regarding negotiations to obtain access to the model. He stressed the proactive nature of government planning, stating, “At a stage where we do not have access to it, we must still move ahead with countermeasures.”
The government has plans to convene talks among relevant ministries and collaborate with the ruling Liberal Democratic Party (LDP) to establish concrete defensive measures. Chief Cabinet Secretary Minoru Kihara has affirmed that the administration is committed to acting swiftly in response to any potential misuse of such advanced technology.
Industry Access and International Concerns
The Mythos model, which was introduced by Anthropic in April, currently has restricted access. It is primarily available to major IT corporations, such as Google LLC, alongside certain financial institutions. However, both public organizations and government bodies are also pursuing opportunities for access.
Experts caution that if the Mythos model were maliciously utilized, it could enable cyberattacks capable of disrupting operations across a wide variety of sectors. Anthropic itself is noted as a U.S. startup co-founded in 2021 by former OpenAI executive Dario Amodei, and is recognized for developing generative AI with an emphasis on safety principles. The company released its conversational AI, Claude, in 2023, adopting training systems designed to prevent harmful behavior and maintain human oversight.
On the global stage, concerns are shared by international leaders. Finance ministers and central bankers from the Group of Seven advanced economies are scheduled to meet in Paris next week (Monday and Tuesday) where they are expected to discuss their collective worries regarding cyberattacks facilitated by cutting-edge AI models, including Mythos.
