Amplifying the Mind with Interactive AI

Imagine a world where the limits of human intelligence are stepping stones to greater potential. By integrating artificial intelligence (AI) systems, we can amplify and expand our cognitive abilities beyond current boundaries. Yet, there is a challenge of effectively assessing and self-assessing the combined performance of humans and AI to fully harness this potential.My project will lay the groundwork for enhancing human intellect through AI integration by addressing three ambitious objectives:(1) making AI-augmented intelligence measurable through comprehensive human-AI assessment; (2) unconstraining metacognition in human-AI interaction to improve self-assessment and empower users;(3) building AI interfaces with a metacognitive interaction model in an ability-centric design process.To achieve these goals, the AmplifAI project will undertake three interlinked work packages. WP1 focuses on conceptualizing and measuring AI-augmented intelligence by developing tasks and metrics to assess human-AI composite performance, conducting large-scale testing, and exploring neural correlates via neuroimaging studies. WP2 aims to understand and remove metacognitive constraints by studying how users assess their performance with AI assistance, using Brain-Computer interfaces to track metacognitive processes, and developing adaptive interfaces to enhance self-awareness in decision-making. WP3 involves creating new AI interfaces with a metacognitive interaction model, iteratively testing prototypes, and launching the AmplifAI Challenge. Ultimately, I envision AI seamlessly integrated into human cognition, maximizing user control and enhancing capabilities. By making both assessment and self-assessment in human-AI interactions possible, we can set the stage for breakthroughs in human-AI synergy and empower people to solve complex global challenges with AI.