Responsible AI
The limited role of AI in Aeviant’s research and scientific communication.
This statement describes AI’s current supporting role in Aeviant’s research. It does not claim controls or governance processes that have not been implemented.
A supporting tool
Aeviant uses AI to assist reasoning, computation, information organization, and the generation of questions for further investigation. AI output is not ground truth and is never represented as experimental evidence.
Verification remains with Aeviant
Scientific claims and citations produced with AI assistance are checked against primary literature. Computational conclusions are evaluated against predefined criteria and adversarial review rather than accepted because a model produced them.
No AI system can advance a program, convert a prediction into a finding, or replace the experimental work needed to test a research hypothesis.
Review is described honestly
AI-assisted criticism can be useful, but it is not a human advisory board. Aeviant identifies automated review as automated and does not present model-generated commentary as independent expert endorsement.
A proportionate public policy
This statement is limited to current scientific use at a founder-led discovery-stage company. It will be revised as Aeviant’s systems, operations, and obligations develop.
Last reviewed .