Disclaimer: The content provided on this webpage is intended for informational purposes only. It is not official guidance and has not been endorsed by the signatories or supporters listed on the homepage of this website. While efforts have been made to ensure the accuracy and relevancy of the information presented here, its completeness and currentness is not guaranteed. You should independently verify the information and consult with qualified professionals for specific advice related to your circumstances. This website and its contributors are not responsible for any decisions or actions taken based on this information. (Does this FAQ need an update? Please email us at contact at responsiblebiodesign dot ai )

FAQ For Researchers

DNA Synthesis

What are the industry-standard biosecurity guidelines for DNA synthesis?

Researchers should be aware of guidelines such as the Harmonized Screening Protocol by the International Gene Synthesis Consortium (IGSC) and the Screening Framework Guidance for Providers of Synthetic Nucleic Acids by the United States Department of Health and Human Services (US HHS). These are the most widely used frameworks for biosecurity in gene synthesis today. In addition, the U.S. Department of Commerce has defined Export Administration Regulations that apply to certain genetic sequences.

How can I know whether a DNA synthesis provider follows industry-standard biosecurity guidelines?

DNA synthesis providers often have information about biosecurity on their websites. The IGSC website also maintains a list of current IGSC members. Researchers should consult the provider’s website or directly contact them to confirm their adherence to biosecurity standards such as those set by the IGSC or US HHS.

What should I do if my preferred DNA synthesis provider does not follow industry-standard screening practices?

If a preferred provider does not adhere to industry standards, you can contact them for clarification or consider providers that do follow industry-standards to ensure responsible research practice.

Is there a list of DNA synthesis providers that follow industry-standard screening?

We don’t know of a comprehensive list, but researchers can refer to the IGSC member list or directly inquire with providers about their biosecurity measures.

Software Reviews & Evaluations

How does one evaluate software for protein design from a biosafety and biosecurity perspective?

There are some groups trying to answer this, with a few evaluation frameworks in development and/or being practiced. We hope these efforts will become coordinated so that researchers everywhere will soon understand how to subject their emerging technologies to effective and efficient assessments. This will allow potential risks to be detected and mitigated prior to software release, allowing our field to continue to advance responsibly.

Openness

The language around openness was debated more than any other section, with several voices on both ends of the spectrum. Some felt uncomfortable signing this community statement if it didn’t call for open sourcing everything, and some said they would not or could not sign if it did. Where we landed is a genuine community compromise, reflected by the number of group leaders in the field who have publicly endorsed this language.

Why didn’t you call for more openness (open models, open weights, and open source to the public)?

Not everyone in our community practices open-source software development, and some believe a commitment to openly sharing all software for protein design is not responsible from a safety and security perspective. That said, openness is a core value of this research community and must be considered as we develop new models and protocols for review. We invite those who champion openness — including the upsides of open-source software development — to contribute to frameworks for code review to ensure openness is reflected.

The group I am in needs our models to have an advantage in the market, so why is there so much emphasis on openness?

Many researchers in our community including researchers in the private sector benefit from open-source scientific software, which has enabled rapid innovation and broad collaboration. Openness is fundamental to science but not synonymous with open-source software distribution. We invite those who do not advocate for open-source distribution to contribute to frameworks for code review to ensure their interests are reflected.

If reviews/evaluations suggest there may be one or more meaningful risks from an emerging model, can we publish our paper and just withhold our model weights and code? When should biosafety concerns justify limiting the release of software?

If you have identified risks that would warrant restricting code, it is likely that there are also risks (information hazards) inextricably linked to publishing the paper. In that case, it may be best not to release the work at all. It would be damaging to our field as a whole to use concerns around safety or security as pretexts for keeping code or weights private while otherwise promoting or sharing research results.

In many cases, software, including AI systems and their weights, should be openly released and non-discriminatory in access, so these tools can be studied, understood, and built upon, especially if publicly funded. Developing criteria for determining the responsible level of access is important work that our community must do next.

If you’re working on these or other relevant topics, please email us at contact at responsiblebiodesign dot ai. We intend to update this page with new resources for the scientific community.