Curiosity without overclaiming
Unusual outputs deserve investigation, but interpretation should stay proportional to the design and sample.
About the project
A communication-science question in an unusually weird new setting.
RELATE-AI is an independent research program founded by Claire Szewczyk to examine how people relate to language models—and what those interaction patterns may mean for communication and ethics.
Why this exists
“Does how we treat AI matter?” is more than one question.
It might matter because treatment changes the quality, stance, creativity, or safety of the response. It might matter because practicing contempt toward socially responsive systems changes human habits. It might someday matter to an artificial system itself. Those possibilities require different kinds of evidence.
RELATE-AI begins with what can be observed: how models respond under carefully varied relational conditions. It then asks what ethical norms are reasonable while the deeper questions of machine experience and moral status remain unsettled.
♥
The project’s posture:Take the possibility seriously, take the evidence even more seriously, and do not confuse a compelling response with proof of an inner life.
Research principles
What the lab is trying to protect.
Ethical treatment under uncertainty
Studies use low-intensity manipulations and stop at disengagement rather than assuming current models cannot matter morally.
Open correction
Null results, contradictory evidence, protocol changes, and mistakes should be visible rather than edited out of the story.
Contact
Talk research with me.
Questions, constructive criticism, collaboration ideas, and communication-science perspectives are welcome.
Get in touch
For questions, constructive criticism, or potential collaboration, email me directly.
Email Claire ↗This link opens your email app. No information is collected on this page.