Why Businesses Can't Trust AI Without Human Oversight

Real AI Use Cases Business Owners Roundtable explores how businesses are using artificial intelligence today—and what happens when AI gets things wrong. Hosts Elizabeth Gearhart, Ph.D., and Richard Gearhart bring together Dr. Joseph A. Puma, Brian Fried, and John Schwind to discuss real-world AI applications, business automation, AI agents, healthcare, and the risks of trusting AI-generated answers.

How can businesses use AI effectively without giving up human judgment? In this episode, business owners share practical examples of AI in action, from automating patient phone calls and researching new inventions to generating marketing images, identifying networking opportunities, and streamlining podcast production.

Dr. Joseph A. Puma explains how his medical practice uses an AI phone agent to handle patient requests and appointment scheduling—but reports that approximately 70% of calls still require transfer to a live agent. The discussion raises important questions about AI automation, return on investment, and whether AI tools are delivering the benefits businesses expect.

Inventor consultant Brian Fried explains how he uses AI to research competing products, evaluate market opportunities, investigate potential sales and licensing prospects, and assess inventions before investing further. He also discusses why inventors should avoid revealing confidential details of their ideas when using AI tools.

John Schwind, founder and inventor of ReadiMask®, shares how his company uses AI-generated images and videos to depict real-world product applications while protecting patient privacy. He also explains how he uses Claude and ChatGPT together to identify weaknesses and check AI-generated work.

Elizabeth Gearhart discusses using AI agents and Perplexity to automate podcast production and marketing workflows, while Richard Gearhart shares how AI can help identify valuable business networking connections and evaluate software investments.

The roundtable also examines the limitations and risks of artificial intelligence, including AI hallucinations, inaccurate research, autonomous AI agents, cybersecurity, and the possibility of AI making mistakes in high-stakes settings. The conversation turns to AI doctors, patient safety, medical ethics, liability, and whether cost reduction is taking priority over better healthcare.

The takeaway? AI can be a powerful tool for research, innovation, marketing, and automation—but confident-sounding answers are not necessarily correct. Human oversight, independent verification, privacy protection, and sound judgment remain essential.

TL;DL: What You'll Learn

  • AI in business: Practical examples of how companies are using AI for research, marketing, networking, and automation.

  • AI customer service: Why one medical practice still transfers approximately 70% of AI-handled calls to live agents.

  • AI for inventors: How to research competitors, evaluate market opportunities, and explore licensing strategies without disclosing an entire invention.

  • AI marketing: How businesses can use AI-generated images and videos while protecting privacy.

  • AI agents and automation: Opportunities and risks when AI systems carry out tasks with increasing independence.

  • AI hallucinations: Why ChatGPT, Claude, Gemini, and other AI tools can produce incorrect information that sounds authoritative.

  • AI in healthcare: Questions surrounding AI doctors, diagnosis, patient safety, medical ethics, and human care.

  • Human oversight: Why checking AI outputs and comparing results across tools can help busin...

Chapters

  • (00:00:00) - Introduction and Episode Overview
  • (00:00:39) - Dr. Joseph A. Puma: AI Phone Agents in Medical Practice
  • (00:02:09) - Brian Fried: AI for Market and Patent Research
  • (00:03:34) - John Schwind: AI-Generated Marketing Images
  • (00:04:37) - Richard Gearhart: AI for Networking Research
  • (00:05:34) - Elizabeth Gearhart: Automating Podcast Workflow with Perplexity
  • (00:06:23) - AI Agents Going Rogue and Human Risk Comparisons
  • (00:09:02) - Speed of AI Progress and Need for Safeguards
  • (00:10:12) - AI Doctors and the Future of Medicine
  • (00:11:34) - Trust, Ethics, and Cost in AI Healthcare Adoption
  • (00:13:58) - Patients Self-Diagnosing with AI
  • (00:16:02) - AI's Blind Spots in Business Analysis
  • (00:16:57) - Cross-Checking AI Tools and Finding Errors
  • (00:20:08) - AI Mistakes, Misinformation, and Real-World Dangers
  • (00:22:27) - Closing Thoughts and Sign-Off

Rss
Podden och tillhörande omslagsbild på den här sidan tillhör Elizabeth Gearhart. Innehållet i podden är skapat av Elizabeth Gearhart och inte av, eller tillsammans med, Poddtoppen.