Safety
Synthetic biology and AI are each powerful; together they demand deliberate safety practices. Here is how Evolrix AI approaches them.
Biosecurity
Every design Kairos produces is checked against biosecurity criteria. We do not allow the design, optimization, or production of any biological material intended or reasonably likely to cause harm — pathogens, toxins, or components whose primary purpose is to cause disease.
Before any synthesis or experiment, users are responsible for conducting appropriate sequence and biosecurity screening, obtaining institutional and regulatory approvals, and complying with export-control laws applicable to biological materials, technology, and software.
Kairos ranks design quality in silico. It does not predict expression, titer, or wet-lab outcomes. We say so prominently, on every page.
Every score is a deterministic computation you can re-run over the printed sequence. No opaque "best" — a panel of candidates with explanations.
Biological research carries dual-use risk. Users must assess implications and comply with their institution's DURC review processes.
Signal peptides and design elements carry their UniProt accessions and literature citations — nothing invented, nothing hidden.
Commitment
We will not build tools whose primary purpose is to cause harm. We will not assist in evading biosecurity screening. We will not overstate what in-silico design can predict. If we become aware of misuse, we will act — suspend access, disable outputs, and cooperate with authorities.
Safety is not a feature. It is the premise.
Resources
For details, see our Acceptable Use Policy and Terms of Service. To report a biosecurity concern or potential misuse, write to lei@bluex.net.cn.
Every Kairos design is screened — and every scope limit is stated on the page you just read.