New APPS: Art, Politics, Philosophy, Science

Epistemic Justice with AI: indicators, diverse training data and knee pain

Knee pain is common and debilitating, and it’s often caused by osteoarthritis in the knee.  Treatment options range from analgesics (including opioids) to knee-replacement surgery.  If you go to the doctor with arthritic knee pain, you can get an x-ray which can then be interpreted using standard rubrics like the Kellgren–Lawrence Grade (KLG) to quantify damage to your knee and then guide treatment options. The KLG isn’t perfect in that the correlation between pain and objective scores of damage to the knee isn’t perfect.  Some people’s knees are a wreck and they report no pain; others have pain beyond what their KLG score indicate. But here’s the thing: Black patients consistently report more knee pain than white patients.  They also tend to have more knee damage on the KLG – but even when you factor that in, Black patients report much more knee pain than white patients with comparable KLG scores.  What’s going on?

One possibility is that factors external to the knee – stress, for example – explain the higher pain.  If that’s the case, then patients need less knee treatment.  But what if their knees were in worse shape?  To answer that question, you’d have to ask yourself what in an x-ray indicated poor knee condition.

Disease is often measured through indicators, and we know that these indicators can lead to all sorts of complexity.  In the context of Covid, for example, there are all sorts of questions about testing and sensitivity that I’ve talked about before.  Along the way, I referred to a fantastic paper on malaria testing in sub-Saharan Africa – suffice it to say that “cases of malaria” reported to donor organizations is a difficult number to parse for reasons having to do with vagaries in testing and diagnosis.

In a new paper in Nature Medicine, a team led by Emma Pierson makes ingenious use of artificial intelligence to tackle the problem of racial disparities in knee pain.  Since algorithms and data are so often implicated in increasing or magnifying racial disparities (see, for example, Safiya Noble on Google, or Timnit Gebru on facial recognition, or Margaret Hu’s chilling “Algorithmic Jim Crow”), it’s encouraging to learn about machine learning working to undermine racial disparities.  Ordinarily, you train an algorithm to perform like an excellent clinician.  In this case, that would mean training it to look at radiography and determine the correct KLG score.  The trick here was to instead train it to look at pain: to determine what features of the x-ray predicted that the patient would report pain.  It turns out that the algorithm’s diagnoses reduced racial disparities in diagnosis by a jaw-dropping 47%.

BIPA Continues to Gain Teeth: Data Retention Policy & Standing

By Gordon Hull

I’ve written about the importance of Illinois’ Biometric Information Privacy Act (BIPA) before (see also here). Briefly, BIPA is the most important and powerful of the (relatively few) state laws designed to protect biometric privacy. The statute establishes a notice-and-consent regime (sigh. better than nothing, though N&C doesn’t work well, and is disturbing as a norm) for private parties that collect biometric information like face scans, establishes the need for data retention policies, establishes a private right of action (individuals can sue; other states make you go through the state attorney general) and establishes a statutory harm – as underlined by the Illinois Supreme Court, violating the statute is enough to collect damages.

Companies like Facebook have been fighting BIPA hard, because it’s bad for their business model, and some of the main litigation has been around Facebook’s phototagging feature. A lot of the issue has been about standing – whether aggrieved parties have the right to sue. Standing sounds simple: in order to have standing under Article 3 of the Constitution, three conditions need to be met. Going through them, however, will indicate why this is harder than it looks.

• First, the plaintiff must have suffered an “injury in fact,” an invasion of a legally protected interest which is (a) concrete and particularized, and (b) actual or imminent, not “conjectural” or “hypothetical.”
• Second, there must be a causal connection between the injury and the conduct complained of, the injury has to be fairly traceable to the challenged action of the defendant, and not the result of the independent action of some third party not before the court.
• Third, it must be "likely," as opposed to merely "speculative," that the injury will be "redressed by a favorable decision."