Aristotle’s insight that virtue lies in moderation is reimagined by Hasan Davulcu and his team at Arizona State University’s School of Computing and Augmented Intelligence. They are developing artificial intelligence systems that prioritize understandability and control, creating what Davulcu calls the “golden mean of AI.” Traditional large language models (LLMs), like ChatGPT, often replicate internet biases, leading to toxic outputs. To counter this, Davulcu’s innovations include tools that demystify AI “black boxes,” allowing users to comprehend and adjust the AI’s decision-making processes. His approach includes recognizing the stance of online conversations, effectively mapping ideologies to encourage balanced discourse. This method significantly reduces harmful outputs—up to 85%—promoting civil interactions. Additionally, a control layer ensures AI applications remain effective across different models, providing users with reliable mechanisms for correction. Overall, these advancements aim to create safer, more transparent AI that upholds shared human values.
Source link
