We’ll begin with several guiding principles for Critical AI Literacy, adapted from the Student Guide to AI Literacy, 7 essential principles. These principles are proactive, non-prescriptive, and can be applied to your own context and needs.
Learn How AI Works
Understand how AI systems work, their mechanics, strengths, and weaknesses. Learn about what data is used to train AI systems and where it comes from. Ask questions, be curious, try things, and share what you know and learn from others.
- Understand that AI is made by powerful private companies with very little to no public oversight. AI companies share little about training models so we don’t know how AI generates its content, leading to a general lack of explainability, reproducibility, and accountability.
- Learn how AI can reproduce bias: AI is trained on datasets containing stereotypes, conspiracies and falsehoods, reinforcing existing biases and potentially causing harm.
- Learn about AI use at UW: Review any statements on AI use in your course syllabi, and consult with professors, research mentors, librarians and peers about expectations around AI use. The UW Student Conduct and Community Standards Office also provides guidance on AI use at UW.
Learn How to Use AI Ethically
Recognize the ethical implications of using and resisting AI in different contexts. Properly disclose how you use AI-generated content. Deepen your critical thinking skills and ability to evaluate AI-generated content and spot false information, biases, and fake images, video and data.
- Consider how AI impacts author rights: By taking and reproducing content without consent, many AI tools infringe on authors’ intellectual property and violate copyright. Consider how AI negatively impacts the livelihood of many job types: artists, authors, programmers, and more.
- Understand that private companies are not transparent about AI energy and resource consumption, and the environmental and human costs of AI are estimated to be high.
- Recognize that AI often “hallucinates” false information and research citations, requiring us to fact check. AI-generated “deep fakes” and hallucinations can also erode public trust.
- Cite or attribute AI-generated content in your work and learn more about Academic Integrity and AI. Here is a guide on how to cite AI for most academic fields and citation styles. You can also check in with your UW professors, research mentors, librarians or peers about AI use and citation expectations.
Prioritize Privacy and Security
Always remember that most AI systems are not private. Anytime you use AI, you have limited or no control over how your personal data will be used. Use only reputable platforms, understand the terms of service, and share as little information as possible about yourself and others.
- Consider how AI may violate personal privacy: AI is trained on our data without our consent, and AI tools often collect our personally identifiable information.
- Before using AI, investigate the security and privacy features of AI tools. Know what they will do with your information and what data is used to train the tool. Make sure you feel comfortable with your input data or prompts being shared if you decide to use it.
- If using AI, learn about the tools that have been vetted by the UW.