All news articles
AI
Lorikeet News Desk
May 16, 2025
TL;DR
AI's role in healthcare is expanding in all directions, but maintaining trust and empathy is crucial for adoption in patient care applications.
Deepali Joshi of Omada Health discusses the need for full transparency and human-centered design in AI use.
Joshi warns against over-reliance, advocating instead for a balance between human judgment and machine support.
The tonality and the layer of empathy—genuine human empathy—cannot be faked; it needs to be built into the system.

Deepali Joshi
Head of Customer Success | Omada Health
Humanity is at the heart of healthcare. As AI takes on a larger role across the industry, the real question isn't what it can do, but whether it can protect the trust, connection, and empathy that makes care truly human.
For Deepali Joshi, Head of Customer Success at Omada Health, transparency and human-centered design are not nice-to-haves, but non-negotiables for any responsible use of technology in care.
Transparent tech: "Empathy and trust are the two key foundations of healthcare," says Joshi—and that applies just as much to technology as it does to bedside manner. When AI tools like LLMs are used in patient interactions, she believes transparency is essential: "It's our responsibility to share that information with the patient and say, 'Hey, this information is coming from an LLM,' or 'Hey, I'm a bot.'" Without that openness, patient trust begins to erode.
Accountability counts: Joshi acknowledges the "fantastic things" LLMs can do—like generating personalized health insights from glucose data—but warns against overreliance. "Ultimately, a provider should be responsible and accountable," she says, emphasizing the need for a thoughtful balance between human judgment and machine support.
Empathy and trust are the two key foundations of healthcare. It’s our responsibility to share that information with the patient and say, ‘Hey, this information is coming from an LLM.’

Deepali Joshi
Head of Customer Success | Omada Health
Empathy by design: For AI to truly support care, Joshi says, "the LLM needs to be very strong. It needs to solve the problem that it is intended to solve for, and not just start hallucinating." But technical reliability isn't enough. "The tonality and the layer of empathy—genuine human empathy—cannot be faked; it needs to be built into the system," says Joshi.
The dark side: Joshi sees AI unlocking a new era of healthcare, powered by vast data insights. But alongside that promise comes risk. She warns of "the dark side," urging leaders to confront data complexity head-on and prioritize trust, privacy, and informed consent.
The deeper challenge, she says, is ensuring AI strengthens—not replaces—the human connection at the heart of medicine. "Could that be possible?" she asks. "Time will tell."