Tether

The policy describes several functions (risk assessment, identity verification, compliance checks, generating insights from aggregated data) that are typically powered by AI/ML in modern financial and compliance systems, especially in the cryptocurrency space. It also lists third-party vendors whose services strongly imply AI/ML use. However, it avoids explicit AI terminology like 'artificial intelligence' or 'machine learning'.

Automated risk assessment, identity verification, and compliance checks for cryptocurrency transactions, anti-money laundering (AML), and counter-terrorist financing (CTF). This includes screening against databases of politically exposed persons and heightened risk individuals.

Processing aggregated user and device data to generate insights, improve services, and optimize the platform's performance and operations.

The policy indicates automated processing for critical functions like identity verification, risk assessment, and compliance checks (AML/CTF). These processes, even if not explicitly 'automated decision-making', heavily influence a user's ability to access and use the service, potentially leading to restrictions or denial of service based on algorithmic flags. This has a direct and significant impact on user access and financial activities.

The 'Appendix - List of Third-Party Service Providers' explicitly names Notabene, Chainalysis, and Refinitiv, detailing their services which involve compliance, risk assessment, and screening – functions commonly powered by AI/ML in the financial sector.

While personal data is aggregated and used for 'improving and enhancing services' and 'generating insights,' the policy states it's 'no longer capable of identifying an individual' when aggregated. It does not explicitly state that personal data *before* aggregation is used for AI/ML training, nor does it explicitly mention 'training' any models.

User interaction data ('patterns of usage') is aggregated for service improvement and insights. However, the policy does not explicitly state that this data is used for *training* AI/ML models, only for 'improving and enhancing' and 'generating insights'.

The policy makes no mention of collecting or using public content for any purpose, including AI training.

Aggregated statistics (derived from user data) may be made publicly available or used by Tether Hadron, but there is no explicit statement about sharing data for *AI training* purposes with third parties. The listed third-party vendors perform services for Tether, but it's not stated that they train their own AI models with Tether's data.