If you read the news about Apple watches from almost a decade ago, you’d be excused for being surprised to read the news today.

Let’s start with CNN in 2015:

…the back of the watch rapidly flashes green and infrared light at your skin, which gets absorbed or reflected by your blood. When your heart beats, there is more blood in your wrist, while there is less blood between beats. By sensing the timing between your heartbeats, the Apple Watch can calculate your heart rate.

As it turns out, solid-colored tattoos — particularly red ones — also absorb the green light and reflect red light. Black tattoos, which absorb both green and red light, can also screw up the Apple Watch’s heart rate sensor.

The Apple Watch works fine with dark-colored skin, scars and skin abrasions, which are translucent, allowing light to pass through. Tattoo ink is opaque, so that light can’t penetrate your skin.

Did you catch the most important line?

The Apple Watch works fine with dark-colored skin…

A new lawsuit (filed Dec 24: PDF) states that this is untrue, has been known for decades to be untrue, and deserves class action again fraud.

Perhaps most notable is that it puts a consumer watch in context of buyers expecting help with COVID-19 survival.

…reliance on pulse oximetry to triage patients and adjust supplemental oxygen levels may place Black patients at increased risk for hypoxemia.

The lawsuit thus brings to light (pun not intended) that structural racism in engineering can end up having very real impact.

We’ve seen this repeatedly before, like a direct line from Kodak photo processing to Google image analysis having the same basic racism.

Apple pretending its products were never meant to be trusted would be the worst tactic here. Everyone knows their data silo strategy expects you to depend on them whole heartedly in every minute of your life.

The fact that someone can prove customers with darker skin should have significantly lower trust in Apple engineering than lighter skin customers… is what will make the “you shouldn’t have trusted us” defense particularly toxic.

Expect lawyers from Apple to say stuff like they told people not to depend on Apple for important health data (should use doctors instead), while engineers and marketing from Apple go on pushing the opposite: only brand you can trust for important data management.

The reality of course is that if you sell an oxygen sensor during a pandemic, you know exactly what people are buying it for and you’d better be treating it as such (whether adding clearer warnings, fixing defects, or pulling it off shelves).

The disaster at Google is a good precedent in this case. Engineers had built image recognition and tested it on human photos they collected. White and Asian faces were embarrassingly misclassified as animals. The engineers corrected the code to prevent this, and released it to the public. Immediately thereafter, people with Black skin complained they were being classified as animals.

Google engineers didn’t test for human experience, they tested for Google engineers’ experience, which exposed systemic American racism.

The lawsuit thus comes at a time when greater awareness of bias in tech has converged with increasingly severe impact from that bias.

To treat the lawsuit properly, Apple should demonstrate anti-racism in its release pipeline as much as it should demonstrate usual engineering requirements like anti-vulnerability (safe development lifecycle).