There has been much speculation over @realDonaldTrump 's health due to a lack of transparency from our gov. I decided to run his face from the last Walter Reed video & the White House steps video against a #deeplearning model of 440 faces of #COVID19 patients who received O2…

due to low O2 sat within a week of initial diagnosis. 1/2 the patients in the dataset died. The President's face was ranked on a scale of 1-7 against a subset of the data (males 69-76 years old). 1 being lowest likelihood of death & 7 being highest likelihood of death...

The size of this subset was 180 patients. Please keep in mind that this is part of an art project and the science is neither proven or reliable. Still, the use of #AI in surfacing risk factors is nothing new. I am an artist, not a medical professional, so...

the result of this intervention should be treated only as art. The President's face ranked a 5, which would indicate a greater than average likelihood of death as a result of COVID. Again, I believe it is important to be clear that despite my outspoken disapproval...

of the President, I do hope that the virus that he helped propagate in the United States through negligence does not take his life or even one more person anywhere for that matter.

PS: I am well-aware that the dataset used here is VERY small and conventional #AI wisdom dictates the necessity of many 1000's of faces to train a quality model. Sometimes in my work I use small datasets & insights gleaned from "overfitting" as a metaphor for #implicitbias

#AI indicates above average likelihood that @realDonaldTrump will die from #covid19? Read this thread.