If reading now about AI TRiSM for the first time, it might be good read on and learn what it’s about because it leads the top 10 technology trends for 2024 according to Gartner.
AI TRiSM stands for AI Trust, RIsk, and Security Management, and is about AI governance. Artificial Intelligence (AI) is currently taking the industry by storm. But how well is AI understood? Once an objective for implementing an AI system has been identified, the next thing that should come to mind is realizing that AI models and applications may not all be totally trustworthy and without risks. It would be wise to put in place a program of governance to ensure that an AI system can be trusted and that the associated risks are mitigated. AI TRiSM helps to design such governance by highlighting general areas of potential concerns with AI.
Confidence in AI models first requires to understand them, as well as the data used with the models. Correct understanding will help to identify and prevent biases. Testing and validation, rigorous and continuous, should be implemented to instill confidence that the models are accurate and fair. A simple first test could be to just try to explain how the models function.
AI systems usually treat with large and important datasets. These systems are likely implemented by using a complex combination of various open source software, commercial software, and online services, with plenty of data movement and data duplication. Are there privacy risks or other risks involved? Understanding, managing, and mitigating the risks is important as well as identifying consequences, which may be of both commercial and regulatory nature.
An AI system that is certainly composed of many components may prove challenging to fully secure. Security controls for every step should be included in an AI governance program.
When building an AI system, putting in place proper governance can help maximize its efficiency and limit the dangers that could come from it. To find out more on AI TriSM please consult this article on Gartner: https://www.gartner.com/en/articles/what-it-takes-to-make-ai-safe-and-effective
Article by Nicolas