Sharkbytes.net

Season 5: Episode 5: Training AI Models with Neurodivergent Human Talent
Season 5: Episode 5: Training AI Models with Neurodivergent Human Talent
A conversation with Peter Kant, Chairman and Chief Executive Officer, Enabled Intelligence.
Having served as an executive at one of the leading tech innovation institutes, SRI International, Pete identified a gap in how AI models were being trained – and thus an opportunity. Without high-quality data inputs, AI models, including LLMs, will underperform or worse. That’s why Enabled Intelligence’s focus has been on ensuring the data used to train and fine-tune AI models is expertly and precisely annotated. What makes their approach unique is the active recruitment of neurodiverse talents that are often overlooked. After all, data analysis and labeling require detail orientation, hyperfocus, puzzle-solving, and diverse ways of thinking. They have built a highly talented and motivated team that focuses on individuals who are neurodivergent, unlocking unique superpower skills and abilities to reduce AI bias, enhance model performance, and create highly reliable solutions.