AI in compliance & risk: 9 real deployments

Operates Jev, an AI model that returns probability scores (nouls) for yes/no decision questions used by customers to make decisions about people, transactions, or content

Reported: An insecure direct object reference bug exposed every noul across all customers, leaking decision outputs about real people, transactions, or content

Rolling out invisible textGrain text watermarks in ChatGPT and Codex outputs for EU users under EU AI Act Article 50, with global API opt-in available

Accenture's Faculty unit embedded inside Anthropic with employee-level access to conduct red-teaming, alignment assessments, and safeguard testing of AI models

Text watermarking deployed on AI model outputs for EU AI Act compliance

HHS uses ChatGPT and other language models to scan at least five years of Single Audit Act compliance records across all 50 states.

Traditional Medicare is running WISeR, a six-state AI prior-authorization pilot using algorithmic denials, which congressional Democrats are moving to overturn via the Congressional Review Act.

HMRC confirmed it uses AI to monitor suspected tax evaders' social media posts, cross-referencing declared income against visible spending on luxury holidays and high-end purchases, restricted to criminal investigations.

Using internal AI tool 'Atlas' to score warehouses on a 100-900 scale for unionization likelihood using hundreds of data points including Reddit posts and internal feedback-survey responses

The US Army allegedly used AI tools to evaluate bids for a $450 million White Sands Missile Range support contract.

Reported: A bid protest alleges the tools hallucinated a weakness in Trax International's proposal; the competing bids differed by $29.4 million.