Blog

The Trust Gap: Why Understanding Artificial Persons Matters Right Now

23/07/2026

RAP Lab Perspectives

Two numbers tell the story of where we are in 2026. Nearly all executives, 97 percent, say their company deployed AI agents in the past year, and Gartner expects 40 percent of enterprise applications to embed task-specific AI agents by the end of this year, up from under 5 percent in 2025. At the same time, 59 percent of Americans say they distrust the companies building these systems, only 16 percent expect AI to benefit society over the next twenty years, and mentions of AI in employee performance reviews are up 240 percent year over year, with sentiment flipping from mostly positive to mostly negative in that same span.

Adoption is outrunning trust. Organizations are handing AI agents real authority, decisions, tasks, direct interaction with staff and customers, faster than anyone is confirming whether people are prepared to work alongside them, rely on them, or believe what they say. That gap is not a communications problem or a policy footnote. It is exactly the question the RAP Lab has been studying for years, and it is why this research matters now more than it has at any point since the lab began.

Why the Workplace Can’t Wait

The shift from AI as a background tool to AI as an active participant, making decisions, managing tasks, talking directly to employees and customers, is the subject of Director Shane Saunderson’s own published research on persuasive robotics: how authority changes the way people respond to a machine’s requests, why directness and familiarity affect whether someone complies with a robot’s ask, and how nonverbal cues shape trust in ways people rarely notice consciously. These aren’t hypothetical concerns. They describe, almost exactly, what’s now playing out at enterprise scale as organizations turn AI agents into something closer to coworkers than tools.

The data backs this up. Gen Z workers already report trusting AI more than their own manager for certain tasks, while two-thirds of executives say their organization has already suffered a data leak or security incident tied to unapproved AI use. Companies are cultivating an “AI elite” of employees who’ve adapted, while worrying openly about the rest of the workforce falling behind. None of this is simply a technology rollout. It’s a live experiment in how much authority, trust, and social standing an artificial agent can hold, run at a scale and speed with almost no precedent, and very little of the underlying psychology has been worked out in advance.

The Other Frontier: Care, Companionship, and Loneliness

The same dynamic is unfolding outside the office. The elder care assistive robotics market is projected to grow from 3.9 billion dollars in 2026 to 9.8 billion by 2033, driven largely by robots designed to reduce loneliness and social isolation among older adults, engaging them in conversation, activities, and daily support. Humanoid companion robots are already deployed in care facilities across Japan, South Korea, China, the US, and Europe, with newer models detecting falls, tracking medication, and holding what developers describe as empathetic dialogue.

This is precisely where Shane’s work with the McMaster Institute for Research on Aging (MIRA) sits, studying telepresence robots as mobility extensions for older adults and evaluating whether intelligent companion devices genuinely reduce loneliness in long-term care, or simply substitute for it. The market is racing ahead on the promise that humanlike design can meaningfully ease isolation. Whether that promise holds, and under what design conditions, is an open empirical question, not a marketing claim, and it’s one the field needs answered before these devices become standard fixtures in how we care for aging populations.

What the RAP Lab Contributes

What sets this moment apart from the usual cycle of technology hype is that the underlying questions are answerable with real methods, not just speculation. The RAP Lab’s interdisciplinary approach, drawing on psychology, artificial intelligence, neuroscience, robotics, sociology, and ethics, exists to answer exactly the questions industry is currently running past: which design choices earn appropriate trust rather than blind trust, when humanlike cues help people and when they’re used to manufacture engagement, and what actually happens, psychologically and organizationally, when a “colleague” or a “companion” turns out to be a machine.

The stakes of getting this wrong are already visible in the data: a public that uses AI constantly while trusting it less every year, a workforce anxious about a transformation nobody fully explained to them, and a booming market for companion robots built on an assumption about loneliness that hasn’t been rigorously tested at the scale it’s now being sold. The stakes of getting it right are just as real. Systems that earn trust appropriately, that communicate their own reliability honestly, and that are designed with a clear understanding of the humans on the other side of the interaction, have the chance to genuinely improve how people work, age, and connect.

The machines are already here, in greater numbers and with greater authority than most predictions anticipated even two years ago. The question this research exists to answer isn’t whether that’s happening. It’s what kind of “person” these systems will turn out to be, and whether we shape that deliberately or find out by accident.

This piece draws on the RAP Lab’s ongoing research into trust, persuasion, and anthropomorphism in human-robot interaction, and on current industry data on AI adoption and elder care robotics. Learn more about our research at raplab.ca.

Sources

Chinonso is a doctoral student researching how technology design shapes human behavior and organizational systems.

Leave a Reply

Your email address will not be published. Required fields are marked *