Certified wet-lab work
Wet-lab work must run in approved BSL-1/2 conditions with traceable protocols and supervision.
Safety & Ethics
Safety is the floor of participation
All teams must follow biosafety, data ethics, academic integrity and responsible-AI requirements throughout registration, experimentation, publication and final submission.
Safety review is a prerequisite for awards
Teams that fail required safety review cannot receive medals or special awards.
Checklist before execution
Safety self-check, material list, risk assessment and emergency plan are required before risky work.
Compliant data use
Sensitive genomic, medical or private datasets require clear authorization and documented handling.
Evidence-based review
Committee decisions rely on submitted evidence, not informal promises or late explanations.
Check-In
Work that requires written approval
The following work requires written approval before experiments or model release.
- Organisms, chassis or DNA parts outside the official whitelist.
- Vertebrates, advanced invertebrates, or their tissue, blood or saliva samples.
- Gene drive technology, antimicrobial-resistance parts or high-pathogenicity pathogen-related sequences.
- Training AI models with large-scale human genomic data, sensitive medical data or other protected datasets.
Before any risky work
Submit the form, material list, protocol, data-source note and containment plan before execution.
Find templates Responsible AI
AI-specific safety clauses
Teams using AI should document data sources, permissions, model scope, risk controls and how outputs are interpreted.
Data transparency
Declare source, licensing, scope, preprocessing and sensitive-data handling for all datasets.
Model safety review
Models that generate biological sequences must include risk assessment and safeguards.
Interpretability
Models guiding experimental decisions must provide basic explanations and verification logic.
Human oversight
AI output cannot replace expert judgment, institutional approval or experimental controls.
Red Lines
Crossing these lines ends participation
No human testing
Never test any synthetic biology product on humans, including team members.
No environmental release
Never take engineered organisms outside approved containment.
No high-risk work
No RG3/RG4 organisms, no BSL-3/4 work, and no unapproved wet-lab activity.
No unauthorized data
Never train or evaluate AI with unauthorized private, medical or protected data.
No harmful generation
Never use AI to generate harmful biological sequences, pathogen designs or operational protocols.
Penalties
Measures scale with risk and evidence
| Level | Measure | Applies to |
|---|---|---|
| Reminder | Correction request or supplemental statement | Minor omissions such as incomplete documentation |
| Warning | Loss of special-award eligibility | General violations such as improper disclosure or promotion |
| Partial disqualification | Disregard offending data or results | Serious issues such as unapproved partial experiments |
| Medal disqualification | No medal awarded | Severe safety or academic-integrity violations |
| Full disqualification | Removed from the competition; public materials may be taken down | Major safety breach or serious misconduct |
Unsure whether your project needs Check-In?
Ask early with materials, data sources, model capabilities and planned containment.