AI scams surge as KuCoin obtains ISO/IEC 42001 certification

KuCoin has obtained ISO/IEC 42001:2023 certification for its artificial intelligence management system. The standard requires organizations to establish and continually improve policies governing their use of AI. KuCoin says the certification covers its AI management system and the organizational functions supporting it. TRM Labs reported in August that criminal AI adoption rose 40% over the past year. TRM's AI-in-Crime Adoption Index scored adoption at 54 out of 100 in 2026. The index scored adoption at 28 in 2024. Scams were the only category classified as Mature. TRM recorded 201 crypto hacks in the first half of 2026. That figure was more than double the previous year. Four percent of incidents accounted for 75% of losses. Infrastructure compromises caused most large breaches. The Digital Omnibus on AI took effect on July 27. It moved obligations for Annex III high-risk systems to December 2, 2027. The annex includes credit-scoring systems. It excludes financial-fraud detection from that category. Transaction-monitoring software therefore does not automatically receive the same high-risk classification. Since August 2, AI transparency rules have required providers of systems intended to interact directly with people to inform users unless the interaction is obvious. The rules include customer-service chatbots. Voluntary certification does not replace financial regulation or data-protection obligations. ISO/IEC 42001 requires organizations to allocate responsibility and monitor AI performance. It also requires organizations to review changes that could alter a model's behavior. BC Wong, chief executive of KuCoin, said greater AI capability must be matched by greater responsibility. The certification forms part of KuCoin's $2 billion Trust Project. KuCoin launched the multi-year program in 2025 to strengthen security, compliance, transparency, user protection, and operational resilience. The certification cannot guarantee that an individual model will always reach the correct conclusion. It assesses whether an organization has processes to govern AI systematically. Future crypto applications could initiate workflows, manage permissions, or execute actions on a user's behalf. Trusted AI in that environment will depend on defined identities, permissions, transaction limits, human-intervention mechanisms, and auditable records.