bias in AI
One example of bias in AI.
What is [answers vary, one example could be training on one biased dataset could lead to biased labels]?
An example of how to make AI more fair
What is diverse and representative data, bias mitigation techniques, diverse development teams, or ethical AI review boards?
One of the 3 main aspects of data rights.
What is the right to knowledge of what and how your data is used, the right to control over your data, and the right to security and privacy of your data?
The definition of feedback loops
What is the system where every biased prediction generates more skewed data, and the system just keeps ratcheting up the bias?
An example of how to make AI more explainable
What is prediction accuracy, traceability, or decision understanding?
Brief overview of the importance of copyright.
What is provides protection and control of works that supports and incentivizes creation?
The definition of human-in-the-loop
What is the system where a humans actively participate in an automated process to provide judgment, feedback, or intervention?
Percent of people who think organizations must be held accountable for their misuse of AI. Closest answer wins [this team gets another turn].
What is 77%?
Explain what complicates the “right to be forgotten” for AI models.
What is the issue where you cannot remove your data from the training of a model?
The three types of data bias
What are sampling bias, historical bias, and measurement bias?
The three "other stuff" responsible AI principles
What are robustness, privacy, and transparency?
One of the requirements for AI output to infringe on copyright.
What is 1. AI model had access to copyrighted work or 2. output is substantially similar to original work