HRDAG To Join the Partnership on AI
Today HRDAG is delighted to announce that we’re joining Partnership on AI to Benefit People and Society (PAI). The Partnership is a multistakeholder organization that brings together academics, researchers, civil society organizations, tech companies and other groups, and HRDAG welcomes the opportunity to engage with all these sectors to better understand the impacts, both beneficial and harmful, of artificial intelligence and machine learning. Our data scientists are looking forward to engaging in discussions that can advance the public’s understanding about AI and to explore best practices for maximizing the benefits and minimizing the harms of these technologies.
“It’s exciting to think about the many ways AI can improve our lives, but at the same time it’s critical that that technologies are developed so that everyone benefits from these breakthroughs. We need to be vigilant, especially in the criminal justice world, about how these technologies might replicate biases and reinforce inequity,” said HRDAG’s lead researcher, Kristian Lum. “HRDAG will work with PAI on how to keep fairness and transparency at the heart of the public conversation.”
“Machine learning and artificial intelligence can in some cases improve our decisions and make our lives more convenient. But if the data used to train the tools are biased, if the data don’t represent the world, then the models will be wrong, sometimes at disastrous cost. We need to ask ‘what happens when we’re wrong? And who bears that cost?'” said Patrick Ball, HRDAG’s director of research. “We need to be sure that we don’t blindly believe computer decisions—we need to think hard about why AI made those decisions, and refuse to use AI when its decisions are likely biased. We’ll be exploring that with PAI.”
HRDAG’s work with PAI will be especially focused on the Partnership’s thematic pillars of Safety-Critical AI; Fair, Transparent, and Accountable AI; Collaborations Between People and AI Systems; and Social and Societal Influences of AI.
- S Prev