Risks & Limitations

AI poses documented risks to student learning, including generation of false information and dependency effects. Broader concerns include bias, intellectual property, the human labor behind data collection, surveillance, environmental impact, and deskilling. 

Learning Loss

Faculty and researchers have identified significant risks of learning loss and deskilling when students become overly reliant on AI tools. Research shows that AI “significantly impacts the loss of human decision-making and makes humans lazy,”  

Yale faculty have observed students forming “illusions of understanding” through AI use in introductory courses, with the illusion “painfully collapsing in upper-level courses when the AI can no longer hand them the answer.” This aligns with broader research indicating concerns about “deskilling due to over-reliance and the potential for impeded skill development due to an over-reliance on technology.”

Key areas of concern include critical thinking, writing and analytical skills, mathematical reasoning, and foundational programming skills. 

AI Ethics

You (and your students) might begin with this overview of AI Ethics with suggestions for how to study and teach about them.

Learn more about AI Ethics

Sustainability

While AI holds promise for smart grids and reducing emissions through increased efficiency, their environmental impact is extreme. Training large AI models like GPT-3 can consume approximately 1,287 megawatt hours of electricity and generate about 552 tons of carbon dioxide. Global data center energy use is rapidly growing.  

To learn more about the environmental and sustainability impacts of AI usage, check out Adam Zewe’s MIT News article, Explained: Generative AI’s environmental impact.

More Resources on AI Risks

Newsletters

Podcasts

Op-Eds

AI Literacy Topics

Explore our resources to learn more.

Circuit board design with the Yale Y tiled through out.