During the Detroit Regional Chamber’s October AI Office Hour, John K. Thompson, lecturer at the University of Michigan, discussed how generative AI can be used to support critical thinking rather than replace it. Presented in partnership with Michigan Online, the session focused on context engineering, AI adoption in the workplace, workforce preparedness, and how organizations can use AI tools while maintaining human oversight and judgment.
Thompson challenged the idea that generative AI diminishes critical thinking, comparing AI tools to other technologies that help people manage routine tasks. He described AI as part of a broader history of cognitive offloading, in which individuals use tools to support activities such as calculations, memory, and editing.
Rather than replacing human thinking, Thompson argued that AI can help people spend more time on higher-value work by automating repetitive tasks and standardizing routine elements of their workflows. As part of that process, he encouraged users to build detailed context documents that outline preferred sources, writing styles, formats, constraints, and expectations for AI-generated outputs. He also explained that providing more context can help improve response quality and reduce hallucinations by giving models clearer guidance about the desired output.
“It allows us to focus on judgment, discernment, ethics, transparency, honesty, all the elements that make us human,” Thompson said. “You really don’t have to do everything in your head.”
A key theme throughout the discussion was the distinction between prompt engineering and context engineering. Thompson described prompt engineering as a single request or instruction, while context engineering incorporates multiple layers of information that guide an AI model’s behavior and outputs.
He encouraged organizations to think about context at several levels, including system prompts, company policies, departmental requirements, process-specific guidance, and individual user preferences. Together, these layers can create a framework that helps AI tools operate more consistently and align with organizational objectives.
During a Q&A session moderated by Megan Spanitz, the Chamber’s Chief Operating Officer and Chief Strategy Officer, Thompson discussed how organizations can prepare employees for continued AI adoption.
He distinguished between AI literacy and AI fluency, describing AI literacy as understanding and being comfortable with AI tools, while AI fluency involves actively applying them to improve productivity and outcomes.
Thompson emphasized that successful AI implementation requires collaboration across organizations rather than assigning responsibility to a single individual.
“AI is a team sport. It’s not something that you ask one person to do,” Thompson said.
He also addressed questions about higher education, workforce development, regulation, and the future of AI in business. While acknowledging that job responsibilities will continue to evolve, Thompson noted that AI will increasingly become part of nearly every profession and industry.
“AI is here. It’s here to stay. Engage with it. It’s going to make your life better,” Thompson said.
