In this episode of the Imagine AI Live Podcast, we sit down with Sheryl Savage, General Counsel & Corporate Secretary of Positron AI—a Nevada-based startup building energy-efficient hardware for AI inference applications. Sheryl shares how Positron’s custom silicon chips are enabling faster, more cost-effective AI processing—ideal for agentic AI, virtual assistants, and large-scale inference use cases. But it’s not just about hardware. As a seasoned attorney, she unpacks the top 5 legal risks for companies using generative AI, including compliance, data privacy, and cybersecurity—and even explains how to use prompts to build your own compliance roadmap. From Limitless AI wearables to the deeper ethics of data collection, this episode covers the real legal, technical, and strategic considerations for building responsibly with AI. (0:00) – Intro and Sheryl’s role at Positron AI (1:31) – Why AI inference speed matters for agents and assistants (2:50) – Sheryl’s 20+ year legal background in tech (3:40) – The 5 biggest legal risks in generative AI (4:46) – How to mitigate privacy and data leakage risk (6:57) – Are we creating new laws or applying old ones? (10:13) – Deepfakes, crime, and the danger of bad inputs (11:40) – Advice for innovators: don’t let legal fear slow you down (13:26) – Where to find Sheryl and Positron AI online