A Trust Governance Layer for AI Agents: Highlights from CR & GliaNet Alliance’s Discussion on Loyalty and Market Signals

Last week in New York City, Consumer Reports (CR) and the GliaNet Alliance hosted a discussion with a small group of technologists, advocates, and nonprofit leaders to explore how we might create effective market pressure for the development and deployment of consumer agents that are truly loyal. We focused on a critical question: as AI agents increasingly have the ability to act on our behalf, how do we enable informed choice and ensure that these agents act in ways that are aligned with our interests?

The discussion built on ongoing work from CR’s research collaboration with Stanford’s Digital Economy Lab, and the GliaNet Fiduciaries Pledge, to imagine governance structures, trust signals, and potential certification schemes for loyal AI systems that represent consumers in the marketplace. Here are some key takeaways and next steps from the workshop:

The Time is Now: Consumer Agency at Risk 

AI agents are rapidly moving from niche tools to everyday intermediaries. Already, they are being positioned to handle shopping, financial decisions, and even post-purchase problem resolution. In this shift, consumers will increasingly depend on agents that claim to act in their best interest. But without shared definitions of loyalty and clear signals of trust, consumers may be left vulnerable to misaligned incentives and hidden conflicts of interest.

The challenge: how do we enable informed choice and ensure agents act in ways that are aligned with consumer interests? 

Our discussion began by posing several questions that we’ll seek to explore in months ahead:

  • How do we define “loyalty” in the context of AI agents?
  • How can deployers build “loyalty by design” into their systems?
  • How can loyalty be measured, incentivized, and evaluated in practice?
  • How can trustworthiness be clearly and credibly signaled in the marketplace?

The Discussion: A Trust Governance Layer for AI Agents 

We wrestled with how to define loyalty. Should agents simply “do no harm” – which is more akin to a general “duty of care” –  or must they actively advance consumer interests – much like a fiduciary duty of loyalty? Some favored a three-tiered model, while others argued for a clearer binary standard. 

Conformance and enforcement was another big topic. We considered who could credibly certify “loyal agents.” Models ranged from CR taking a leading role, to founding a new professional association, to creating a B Corp–style movement. Whatever the structure and process, leveraging existing technical standards from organizations like IEEE, rather than inventing new ones, was viewed as a pragmatic way forward.

This is a mockup by Ocupop, showing how a certification scheme could signal trust to consumers and incentivize businesses

The business model question was front and center too. One important premise is that developing a trust-based commercial relationship between for-profit companies and their customers yields important financial and brand advantag

Lastly, we acknowledged the need for broader collaboration. Future sessions should bring in legal experts, industry leaders, investors, and consumer advocates to test assumptions and build momentum for a certification model that could scale. 

The Road Ahead: Exploring Interest Across Stakeholders 

The workshop concluded with five next steps: 

    1. Map existing technical protocols and standards to new normative definitions of loyalty
    2. Develop a comprehensive “handbook” that grounds fiduciary duties of care and loyalty in actual business practices
    3. Conduct research on effective types of trustmarks for audiences of consumers
    4. Keep the conversation alive in shared spaces
    5. Convene a larger follow-on session with a wider mix of stakeholders

We’re interested in understanding whether there could be real appetite for this work. Would consumers and businesses find value in a “Loyal Agent” trustmark that signals trust? We’d love your perspective. If you’re exploring what it would take to create loyal agents, experimenting with trust signals in AI, or simply curious about how a consumer-facing accreditation scheme might work, we’d love to hear from you. Reach out to the CR & the GliaNet Alliance team at innovationlab@cr.consumer.org

Many thanks to our partners at the GliaNet Alliance and Ocupop who helped facilitate this fruitful discussion. 

Get the latest on Innovation at Consumer Reports

Sign up to stay informed

We care about the protection of your data. Read our Privacy Policy