Issue Home bioengineering enabled by machine learning and kits

Summary

Among longer-term uncertainties, the report raises the prospect of individuals making bioengineered products at home. This could become possible if machine learning that predicts protein structures and drug effects from DNA sequences is combined with home fermentation kits and engineering biology tools. The report holds that open-source resources, cheaper tools, large language models and platforms like DIYbio have already lowered the barriers to these technologies. The wider access becomes, the greater the risk that engineered organisms are released, whether on purpose or by accident. The report also worries that decentralised production could make enforcement harder and lead to fake medicines or synthetic opioids.