Extract from Laura Land Himelstein’s article “Categorizing AI Tools to Calibrate Contract Terms and Third-Party Risk Management Controls”
To effectively leverage artificial intelligence (AI), organizations must navigate a range of complex considerations, including data governance, ethical use, operational deployment, business continuity, intellectual property, privacy, cybersecurity, and more.
As a first step to address these challenges, this article recommends classifying AI tools into three broad categories:
(1) “Deterministic” AI tools utilizing rules-based reasoning, such as robotic process automation (RPA) and machine learning;
(2) “Probabilistic” AI tools using probabilistic reasoning, such as generative artificial intelligence (GenAI); and
(3) “Agentic” AI tools designed to operate with greater autonomy and decision-making capability.
Proper categorization facilitates selection of relevant contract terms and risk management controls, which may become increasingly stringent as AI tools move from deterministic to probabilistic to agentic.
