introduced a broad framework governing AI, categorizing AI tools into four levels based on associated risks: minimal, limited, high and unacceptable. The higher the risk, the stricter the rules.
Suspected Trump Shooter Remembered By Rifle Team Member As ‘Comically Bad Shot’: What We Know About Thomas Matthew CrooksThe limited risk category includes AI systems that interact directly with users, generate content, conduct emotion recognition or biometric categorization or create deepfakes. All of them require transparency. The tools categorized as “minimal risk” include, for example, spam filters and video games.
Be sure to write your company’s policy on AI using simple sentences with clear meanings. For example, in the opening paragraph of our policy, we state,"AI tools should be used to improve your work by providing support and adding value, not replacing the work you do." Your policy also provides criteria for adding an AI tool to the list of approved tools. You should also include a table or section that instructs employees on how to use AI tools within specific company processes. Your policy should encourage employees to use common sense, define areas of responsibility and name specific people who must be contacted if a particular problem arises.