A Quest for Fair AI: The Future Impact of Data Collection

Description:

AI systems, trained on human-generated and web-scraped data, often inherit and amplify societal biases, posing significant risks to fairness and equality. This raises a crucial dilemma: should AI reflect our current society, including its inequalities, or strive to be an unbiased ideal? Who should be making the value judgments that are central to the future of this technology? In this session, panelists will share their perspectives on this issue and suggest ways to improve data collection to avoid perpetuating and worsening existing inequalities while ensuring real-world applicability.