EP 422: Licensing AI Agents – What is it? And Do We Need It?

Podcast: Everyday AI Podcast
Host: Jordan Wilson
Guest: Dr. Denise Turley (Vice President, Technology & Educator)
Date: December 13, 2024

Episode Overview

This episode delves into the emerging debate around licensing AI agents—whether AI-powered virtual assistants and decision-makers should be subject to similar certification, safeguarding, and oversight as human professionals. Host Jordan Wilson is joined by Dr. Denise Turley to explore the risks, practicalities, and societal changes such a shift could bring, particularly as agentic AI becomes an increasingly integral part of our work and personal lives.

Key Discussion Points & Insights

1. What Does "Licensing AI Agents" Mean?

[05:23]

• The idea draws from how many U.S. professions (doctors, pilots, CPAs, beauticians, etc.) require licensing because of the direct impact their work has on human lives.
• Dr. Turley explains the concept:
  • AI agents are growing more capable, autonomous, and engaged in decision-making.
  • Should we, therefore, require them to be tested, certified, and held accountable for their advice or actions, similar to licensed professionals?
  • Quote:
    "If I already have a license, that license extends to my agent... Or we just license the AI agent as technology that's actually giving advice and is qualified to do so, has been tested, has been trained and has ongoing tests and compliance so that you can increase trust and also have accountability." — Dr. Denise Turley [06:32]

2. The Duality of Empowering AI Agents

[09:25]

• Excitement: Licensed AI could accelerate access to services, improve output, and fill gaps in overburdened sectors (e.g., medicine, finance).
• Caution: Licensing suggests granting agency and responsibility—a potential source of fear or skepticism.
• The goal: Achieve a balance, allowing innovation while maintaining safety and public trust.

3. Practical Models of Licensing

[08:23]

• Licensing could be extended from professionals to their AI agents, meaning those professionals remain accountable for their AI’s actions.
• This would require rigorous validation and continual oversight from the humans deploying these tools.
• Quote:
  “If I'm already licensed... there's a framework where I'm allowed to extend that license to my AI agent, meaning that it's acting on my behalf... and that also means, so that I'm now accountable for everything it does.” — Dr. Denise Turley [08:28]

4. AI Agents as a Workforce & Shifting Relationships

[12:53]

• Increasingly, businesses use AI as practical “members” of teams (e.g., meeting note-takers, virtual avatars representing humans in meetings).
• Dr. Turley describes sending her own avatar to meetings, raising questions about future workplace interactions.
• As AI agents proliferate, management and oversight will look fundamentally different.

5. Risks and Dangers of Licensed AI Agents

[15:28]

• Bias in Data: AI may perpetuate or amplify pre-existing biases, with real-world consequences (e.g., healthcare for disadvantaged groups).
• Accuracy & Hallucination: AI confidently produces convincing but sometimes inaccurate or fabricated information.
• Recommendation: Even with licensing, AI advice should be reviewed and validated by a human, especially in high-stakes fields like medicine.
• Quote:
  “We know that the AI hallucinates. We know that it makes stuff up in a way that is absolutely convincing. So we have to figure out where it stops.” — Dr. Denise Turley [16:38]

6. Who Should Oversee Licensing?

[17:51]

• Should government agencies, professional bodies, or entirely new entities supervise and audit AI agent licensing?
• Turley suggests existing licensing organizations could evolve their practices, incorporating technical standards and ongoing auditing.
• There is job creation potential in building new oversight systems for AI.

7. Re-licensing Frequency and Keeping Up with AI Evolution

[20:07]

• Audience question: How often should AI agents be re-licensed, given their rapid learning and updates?
• Dr. Turley: At a minimum, annually, but likely more frequently (potentially daily) as roles, stakes, and technology change.
• Quote:
  “...there's this notion now of what happens when the AI gets smarter and smarter... we would have to then start, I think, having more consistent evaluation of its output. And that might be daily.” — Dr. Denise Turley [21:05]

8. Implications for Human Roles & Choice

[22:23]

• As AI agents become more capable, humans may shift toward oversight and orchestration of multiple AI “staff.”
• Will people prefer to interact with AI or humans in essential services? Choice remains vital.
• Quote:
  "Humans should be able to choose if they want to be treated or engaged with AI." — Dr. Denise Turley [24:02]

9. Personal Preferences and Human Connection

[25:03]

• Dr. Turley: AI is valuable for efficiency and handling undesired tasks but cannot replace the need for genuine human connection.
• Preferences for AI vs. human interactions will be personal—context and trust play key roles.

10. Accountability and Cautionary Tales

[27:08]

• Host Jordan Wilson cites lawsuits involving AI (e.g., Character AI) as a warning: real harm can result if oversight is lacking.
• Both speakers underline the need for robust accountability structures to safeguard society.

Notable Quotes & Memorable Moments

• “If I'm already licensed... there's a framework where I'm allowed to extend that license to my AI agent, meaning that it's acting on my behalf... and that also means, so that I'm now accountable for everything it does.”
  – Dr. Denise Turley [08:28]

• “We know that the AI hallucinates. We know that it makes stuff up in a way that is absolutely convincing. So we have to figure out where it stops.”
  – Dr. Denise Turley [16:38]

• "There is a time and a place for AI, but AI doesn't replace humans... Human connection is important and AI cannot replace that."
  – Dr. Denise Turley [25:03]

• “It's a matter of balancing innovation with safety. Ultimately, we have to keep humans first.”
  – Dr. Denise Turley [28:06]

Key Timestamps

• [05:23] – Defining "licensing" for AI agents
• [08:23] – Extending professional licenses to AI agents
• [09:25] – Host's apprehension and excitement about licensed AI
• [12:53] – Business implications: AI agents as workforce members, meeting avatars
• [15:28] – Primary risks: bias, accuracy, and safety
• [17:51] – Oversight responsibility: who should license AI agents?
• [20:07] – How often should AI be re-licensed? The challenge of “keeping up”
• [22:23] – The evolving human role: from doers to AI orchestrators
• [24:02] – Preserving choice: do users want to interact with humans or AI?
• [25:03] – Personal boundaries and the irreplaceability of human relationships
• [27:08] – The necessity of responsibility and accountability in AI deployment
• [28:06] – Final takeaway: “balance innovation with safety”

Takeaway

Licensing AI agents is fundamentally about safeguarding society while driving innovation. It means building accountability, trust, and human-centered oversight into an AI-powered future—where both choice and responsibility must remain at the forefront.

For additional insights and resources, subscribe to the Everyday AI Podcast and daily newsletter at youreverydayai.com.