AI Copyright and Training Worksheet (Questions 7-10)

What are the three key resources AI companies need to build their models?

• People, data, and funding

• People, GPU, and data

• Engineers, artists, and data

• Software, hardware, and data

How much can AI companies spend per engineer?

• $500,000

• $1 million

• $2 million

• $100,000

What percentage of large language models were found to be trained on copyrighted works?

• 21%

• 64%

• 50%

• 75%

What is a major consequence of generative AI training on copyrighted work?

• It improves the quality of human-created work

• It creates new job opportunities

• It competes with the original creators

• It has no significant impact

What legal argument do many AI companies use to justify unlicensed training?

• Fair use copyright exception

• Public domain argument

• Creative commons licensing

• No legal framework exists

What percentage of people believe AI companies should compensate data providers?

• 50%

• 74%

How did Kelly McKernan's income change after their work was used for AI training?

• It increased by 33%

• It fell by 33%

• It remained the same

• It doubled

What is one solution proposed for AI companies to avoid copyright issues?

• Ignoring copyright laws

• Licensing their training data

• Using only public domain data

• Training on human-created works only

How many lawsuits are currently ongoing against AI companies regarding copyright?

• 10

• 30

• 50

• 100

What does the "Statement on AI Training" address?

• The benefits of unlicensed training

• The threat to creators from unlicensed training

• The advantages of AI training

• The importance of human creativity