Seattle startup uses clinical expertise to make AI models safer and reduce dangerous responses
Seattle-based startup mpathic is expanding to help foundational model developers and LLM-powered application teams reduce dangerous responses and better protect vulnerable users.
Founded in 2021 to bring more empathy to corporate communication, mpathic now supports AI teams by producing evaluation and training datasets designed to make models safer for users in crisis, including children and people with mental health challenges. “We are essentially producing eval sets or training data sets to make models more safe for vulnerable users, like kids or people with mental health problems, people in crisis,” said co-founder and CEO Grin Lord, a board-certified psychologist and NLP researcher. In one early engagement, the company said its clinician-led program helped a model builder reduce undesired or dangerous responses by more than 70%.
As AI tools increasingly serve as first-line interfaces for mental health and medical support, mpathic’s growth highlights the region’s leadership in developing responsible AI technologies that strengthen the life sciences and healthcare innovation ecosystem.