472 karma · joined February 28, 2011
[ my public key: https://keybase.io/heliostatic; my proof: https://keybase.io/heliostatic/sigs/rvCL5g5MnNu8e6ZoO4vrChAihsiizm01niVQ7_2OEC8 ]
Mary Anne Franks
Mike Godwin
James Grimmelmann
Gus Hurwitz
Jeff Kosseff
Emma Llanso
Alan Rozenshtein
Eugene Volokh
Benjamin Wittes
Jonathan Zittrain
Moderated by Kate Klonick
You can contribute to improving it here, too: https://commonvoice.mozilla.org/en
> In less than 6 hours after starting on our in-house server, our model generated 40,000 molecules that scored within our desired threshold. In the process, the AI designed not only VX, but also many other known chemical warfare agents that we identified through visual confirmation with structures in public chemistry databases. Many new molecules were also designed that looked equally plausible. These new molecules were predicted to be more toxic, based on the predicted LD50 values, than publicly known chemical warfare agents (Fig. 1). This was unexpected because the datasets we used for training the AI did not include these nerve agents. The virtual molecules even occupied a region of molecular property space that was entirely separate from the many thousands of molecules in the organism-specific LD50 model, which comprises mainly pesticides, environmental toxins and drugs (Fig. 1). By inverting the use of our machine learning models, we had transformed our innocuous generative model from a helpful tool of medicine to a generator of likely deadly molecules.
Such a good insight -- seems obvious, but too often the source of gotchas, bad data, and bad user experience.