Machine and Human Intelligence: Algorithms as a Source of Bias or Insight

Algorithms can be useful aids in medicine, but they are not foolproof and without bias. This presentation uses three “stories” to illustrate broader lessons for AL and medicine. The key takeaways from these compelling stories is that 1) Data not algorithms are the scarce resource; 2) AI breaks because the data is broken; 3) unrepresentative data can lead to biased results; and, 4) data can be used appropriately for prediction, but not emulation.