Bloom's stAIrcase

What is Evaluate?

Evaluation involves making judgments based on criteria and standards. When we evaluate, we assess the quality, effectiveness, and/or value of something. Evaluation requires critical thinking, reasoning, and justifying conclusions with evidence.

In the revised Bloom's Taxonomy, Evaluate is one of the highest-order cognitive processes. Analyze involves breaking things down to understand them, Evaluate requires making judgments about their worth, quality, or effectiveness. Evaluation enables us to assess arguments, judge solutions, critique methods, and make informed choices.

As a higher-order thinking skill Evaluate requires things like remembering relevant information, understanding concepts, applying criteria, and analyzing components before making judgments. Evaluation differs from personal opinion—it requires systematic assessment using appropriate criteria and evidence-based reasoning. Two cognitive processes within evaluation are checking and critiquing.

AI Literacy Connection

AI can check for certain types of errors and compare against explicit criteria; without human intervention, AI can lack nuanced judgment, contextual evaluation, and ethical assessment. AI evaluation lacks the human judgment, values consideration, and contextual understanding that characterize sophisticated evaluative thinking.

Honing AI Literacy and human-AI collaboration via evaluation might look like, using AI to perform initial checking (grammar, calculations, consistency), then experiment with humanly intervened evaluation; evaluate AI outputs according to a lens or specific criteria; think about the limitations to AI’s ability to evaluate, what human-centered evaluation skills is AI incapable of; evaluate the sustainability of AI use within a particular field; experiment with different levels of human intervention and prompting combined with AI evaluation, assess each iteration of this; evaluate AI tools themselves—their reliability, biases, and appropriate uses; developing the ability to judge when AI evaluation is sufficient vs. when human judgment is needed.