Security & Misuse

An article promotes improving digital literacy by using ChatGPT to verify suspicious emails and SMS as special fraud cases surge.

• •According to National Police Agency statistics, special fraud damages from January to August 2026 totaled 236.07 billion yen, a 58.84 billion yen increase from the same period last year.
• •In the first half of 2026, people in their 60s had the highest damages and case counts, with those aged 50 and older accounting for about 80% of damages.
• •The article explains how to use ChatGPT to help determine if suspicious emails or SMS are fraudulent.

TWOSTONE Sync partnered with Layer8 to launch AI-powered penetration testing services for web applications and APIs.

• •TWOSTONE Sync Inc. formed a business alliance with Layer8 Inc., a developer of AI-based web security technologies.
• •As its first entry into cybersecurity, it began offering penetration testing for web applications and APIs.
• •The partnership enables security services that combine AI and human verification.

A US man was sentenced to prison for fraudulently obtaining royalties by mass-playing self-made AI music via bots.

• •A 54-year-old man in North Carolina used bots to play AI music over 80 million times.
• •He is alleged to have defrauded copyright royalties through this behavior.
• •The man was sentenced to 18 months in prison.

Amid frequent unauthorized access incidents in Japan and views that attackers are abusing AI, this article examines countermeasures to protect Japanese information through interviews with the National Cyber Directorate.

• •Unauthorized access incidents are frequent in Japan, with views that attackers are abusing AI.
• •The article includes direct interviews with the National Cyber Directorate on how to protect Japanese information.
• •The content verifies how AI abuse is involved in cyberattacks.

Anthropic has launched OSS Scanner, a new service that helps detect vulnerabilities in open-source projects by providing free, regular security scans using its top-tier models.

• •Anthropic offers support for tracking security vulnerabilities in open-source projects through its new OSS Scanner service.
• •Opt-in open-source projects receive free, thorough, and regular security scans powered by Anthropic's top-tier models.
• •The service allows open-source projects to potentially receive earlier alerts about potential security issues.

Ethereum co-founder Vitalik Buterin agreed with researcher Justin Drake's warning that AI progress threatens the cryptographic algorithms protecting wallets.

• •Ethereum co-founder Vitalik Buterin agreed with Ethereum researcher Justin Drake's warning that AI progress threatens the cryptographic algorithms protecting wallets.
• •New concerns regarding cryptocurrency security are emerging as AI makes rapid mathematical progress.
• •The warning suggests that improved AI capabilities may impact the security of traditional cryptographic techniques.

AnthropicがClaudeの乱用禁止や選挙介入、偽情報キャンペーンなどを禁止する利用ポリシーを更新した

• •Anthropicは極端なケースでのClaudeの繰り返し乱用を明確に禁止する利用ポリシーを更新した
• •通常の苛立ちや批判は引き続き許可されている
• •新しいルールでは選挙介入、欺瞞的なキャンペーン、兵器ソフトウェア、監視も対象としている

GoodfireがAIエージェントの監視手法を一新し、モデル内部を監視することでコストを大幅に削減した新技術を発表した

• •従来の手法とは異なり、エージェントの全動作を別のAIで読み取らせるのではなく、モデルが動作中に内部を覗き込む方式を採用している
• •異常が疑われる場合のみバックアップを呼び出すことで、監視コストを従来の一部に抑える設計となっている
• •AIエージェントの暴走や誤動作を低コストで検知するための新しい監視ソリューションとして提供される

GoogleがAI生成コンテンツの電子透かしを検出するWebサイト「SynthID Detector」を一般向けに公開し、複数企業のAIに対応した

• •Googleは10月7日、画像・動画・音声の電子透かしを検出するWebサイト「SynthID Detector」の利用対象を一般に広げた
• •Googleのほか、OpenAI、NVIDIA、KakaoのAIで作られたコンテンツに対応している
• •AppleのAIで作られたコンテンツにも近く対応する予定とされている

As AI use by attackers becomes certain amid frequent data leaks, major IT vendors were interviewed on how defenders should utilize AI.

• •Reports of security incidents caused by third-party cyberattacks remain constant.
• •It is widely accepted that attackers are definitely using AI.
• •Major IT vendors were asked about trends and countermeasures regarding how defenders should use AI.