Can we use AI with students?

• We are unable to use most AI tools in the classroom with student logins as most tools will not sign the CA Student Data Privacy Agreement. Per Ed Code, we must have that in place to use student logins with Ed Tech tools including AI tools. 

How can we teach responsible use of AI?

• Students must understand AI in order to develop AI Literacy (The skills needed to understand, use, and evaluate how and when it’s appropriate to use AI)
• We must provide experiences to students: show them bias, hallucinations, misinformation, etc.
• We must provide guidance on evaluating AI, choosing the right tool and/or the right situation when using AI, etc.

Help them understand how AI works

• Teach about data sets, how to train AI, and even look for bias.
  • Oceans AI from Code.org
  • Google's Quick Draw
  • Google's Teachable Machine
• Give them experiences in critically looking at media:
  • Say What You See - Learn the art of image prompting with the help of Google AI.
  • Odd One Out - Can you spot the odd one out? Guess the AI generated “imposters” hidden among the artworks on Google Arts & Culture.
  • Which Face is Real
  • Padlet's AI Tool "I Can't Draw"
  • Rumor Guard from News Lit Nation

Teach about the drawbacks of AI:

• Potential bias in the AI training data: acknowledge the issue and be aware of the potential and help our students learn to review and verify AI-generated content. Here is a great website from Bloomberg Humans are Biased. Generative AI is even worse or the MIT Media Lab's Gender Shades Project
• Lead to over-reliance: AI should complement your decision-making rather than replace it. Use AI as a thought partner and maintain your oversight. Choose when and how you use it carefully and help our students understand this process.
• AI Hallucination: This is when AI "lies" or makes up information. We must teach students to evaluate, verify, cross-check, and not take the AI generated info at face-value. Here is a great example of AI halucinating from the Ventura County Star: We asked ChatGPT to write this story for us. Here's what it came up with.
• Use of AI to create misinformation or disinformation on social media. View the MIT Media Lab's Detect DeepFakes: How to counteract misinformation created by AI or Detecting Deepfakes: Don't Be Fooled by Disinformation | Be MediaWise from PBS Learning Media

AI Powered Scams

Perhaps the most important information we can help our students understand is the way AI is now being used in scams and phishing. Here are just a few ways scammers are using AI:

• Write more convincing and natural-sounding pishing emails and text messages.
• Create deepfakes of celebrities to trick victims into thinking they're investing in a good company or project.
• Impersonate the victim's friend or relative and ask for money as part of a grandparent scam.
• Impersonate an employer and ask for personal information.