I received an email from MyHeritage about their v2.5 ethnicity estimate update.
I logged in, saw a button asking if I wanted to update my results, and clicked it. A message popped up: Check back tomorrow.
So I did. And what I found? The worst update yet.
MyHeritage erased all of my European ancestry. Every last bit.
Let me break this down:
- I uploaded my AncestryDNA results to MyHeritage years ago.
- AncestryDNA correctly detects my European ancestry. My great-grandmother was European—confirmed by her son’s DNA test and thousands of DNA matches.
- Before this update, MyHeritage recognized 22.3% European ancestry.
- I uploaded my DNA to other sites. They all recognized my European ancestry.
- I have Caucasian DNA matches on MyHeritage—people who took their test through MyHeritage. Some also tested with Ancestry, and they have zero African DNA.
- This isn’t a glitch. I’ve seen Caucasian users sharing their v2.5 European results without issue.
- I’ve categorized over 23,000 DNA matches on Ancestry.
- 8,813 of them have detailed notes. I know my DNA matches inside and out.
So what did MyHeritage’s algorithm do? It dumped everything into Nigeria.
That’s not just an error. That’s algorithmic bias.
This isn’t the first time I’ve seen companies do this. DNA tests aren’t perfect—there will always be unknown variables. But instead of acknowledging that, they shove those unknowns into pre-selected categories. If you’re a person of color, they push the “unknown” into Africa. If you’re Caucasian, they probably dump it into a random European country.
I don’t agree with this practice.
If an ethnicity estimate isn’t certain, say that. Give users an “unknown” percentage. As science improves, those gaps will fill in naturally.
But this? This isn’t an improvement. It’s not even neutral. It’s flat-out wrong.
And MyHeritage should be held accountable for it.
