altLabs is dedicated to the safe and beneficial development and use of emerging technologies. We consider the risks that key technologies, like artificial intelligence and biotechnology, may pose to current and future generations. We develop strategies and policies to promote responsible technological development, as well as directly design, build and test technologies ourselves.
Genetic Engineering Attribution
Genetic engineers modify an organism’s DNA to alter its function. This process underpins many scientific discoveries, improving our knowledge of the natural world and guiding the production of new medicines and more sustainable manufacturing techniques. At the moment, however, if you were to stumble upon a genetically modified organism, there would be no way to reliably determine where it came from or who engineered it. Policymakers call this the attribution problem. Without solving attribution there is anonymity. Without adequate attribution, some might be tempted to dodge accountability for their actions, neglecting to engage and get consent from the communities affected by their work, or even deliberately using genetic engineering to cause harm.
We believe that developing technologies which improve accountability and transparency in genetic engineering will help realize the positive impact of this technology while improving safety and security. That’s why we are working to build computational tools, powered by machine learning, to trace the DNA sequence of genetically engineered organisms back to the responsible laboratory, which we call “Genetic Engineering Attribution.”
Working with researchers from Harvard, MIT and elsewhere, altLabs researchers have recently published state-of-the-art models for predicting the lab of origin of engineered DNA. To improve on these results even further, altLabs has partnered with DrivenData to host the Genetic Engineering Attribution Challenge, the results of which will be announced in the near future.
Diagnostics are molecular probes which detect the presence or absence of a molecular marker associated with disease. As diagnostic technology improves, tests achieve higher accuracy, greater convenience, and lower cost. However, these tests still need to know what they are looking for. In other words, because they are designed to detect a known marker, they can only diagnose diseases we have already seen. If we are to be prepared for a new infectious disease, we will need to be able to detect and respond rapidly to something we have not seen before.
To make this happen, we will need to rethink how we approach diagnostics. We will also need to develop a suite of new technologies that can measure the right biological data and algorithms which can process that data into an accurate result. In this project, we are evaluating approaches to this problem, as well as mapping the landscape of technologies which might help us achieve early detection of new pathogens.