Human Judgement. By Design.
Orange Gate Labs helps organizations put AI to work without losing the human judgement that matters.
How I Help
Make good decisions about AI.
Where should AI be used? Where shouldn't it? How do we know it's working as intended. And whether human oversight is meaningful?
I help organizations work through these questions, identify opportunities worth pursuing, and design practical ways to use AI while keeping human judgement in the loop.
Featured Work
What Makes AI So Believable?
Conversational AI can sound remarkably convincing, even when it's wrong. Understanding what makes it believable is one step toward making better decisions about how we use it.
The Believability Effect Cue Detector makes some of those influences visible. Built from original research into conversational AI and human judgement, it highlights the subtle textual cues that can shape how we perceive and trust AI-generated responses.
It's a small intervention designed to spark a bigger conversation about how we engage with AI, and how we might do it better.
Discover
Inside Orange Gate Labs
Orange Gate Labs brings together hands-on advisory work, practical experimentation and writing about AI and human judgement. Each is a different way of tackling the same challenge: making AI useful without losing sight of the people who use it.
Services
I work alongside teams developing and implementing AI-enabled solutions, helping navigate obstacles, challenge assumptions, and move promising ideas forward.
Writing
Essays and observations on AI, human judgement and the messy realities of innovation.
The Lab
Tools, experiments and interventions exploring how people interact with AI and what happens to human judgement.
AI Accidents
Stories of Machine Mischief and Misplaced Trust.
Fluent language can make AI errors surprisingly persuasive. The BE Cue Detector shows how.
The AI Accidents Rogues’ Gallery features real examples of how conversational AI can mislead, confuse, or quietly reshape our thinking.
Each case is a small lesson in how humans judge AI.