All vulnerabilities
CRITICALOpSeccurated

OPSEC-23ANDME-2023

Consumer/genomics · 23andMe

Summary

23andMe held the most personal data there is: people's DNA. In 2023 attackers got into more than 18,000 accounts and, through a single social feature, turned that into the genetic and ancestry data of roughly 6.9 million people. The break-in required no flaw in 23andMe at all. Attackers simply took username-and-password pairs leaked from other companies' breaches and tried them, betting, correctly, that people reuse passwords. The accounts had no MFA, and 23andMe did not notice the five-month wave of automated logins. From those footholds, the attackers scraped relatives' data through an opt-in feature, and the fallout, fines, a $50 million settlement, and ultimately bankruptcy and a fire-sale of the DNA database itself, shows that a breach can be fatal even when your own systems were never hacked.

How it happened

This was a credential-stuffing campaign that ran from late April to September 2023, undetected for about five months. The attacker took username and password pairs leaked from unrelated prior breaches and tried them, by the million, against 23andMe's login. Because so many people reuse passwords, more than 18,000 accounts were directly compromised (23andMe's first, US-focused disclosure put it near 14,000; the later joint UK-Canada regulatory investigation established the global figure at 18,222). 23andMe's own systems were never breached, but it failed to detect or throttle the flood, even when one account was hit with around a million login attempts in a single day, and it did not enforce MFA, with only about 22% of users having it enabled.

Then came the amplifier. From those footholds, the attacker abused the opt-in DNA Relatives and Family Tree features, which are designed to show you information about your genetic relatives, to scrape data on approximately 6.9 million additional people (around 5.5 million through DNA Relatives and 1.4 million through Family Tree), including names and ancestry estimates. Curated lists, for example of people with Ashkenazi Jewish or Chinese ancestry, were advertised for sale, a genuinely chilling targeting risk. One feature meant for family discovery turned 18,000 break-ins into a millions-strong data leak.

The damage

Around 6.9 million people had genetic-relative data exposed, none of whom were among the directly compromised accounts. The ethnicity-targeted lists raised real safety fears. The corporate fallout was severe and lasting: the UK's data regulator fined the company £2.31 million in 2025 for failing to protect genetic (special-category) data, a US class action settled for $50 million (approved in early 2026), and in March 2025 the company filed for Chapter 11 bankruptcy. The data outlived the company: after a contested auction, 23andMe's assets, including its vast DNA database, were bought for $305 million by TTAM, a nonprofit founded by co-founder Anne Wojcicki, a uniquely modern lesson in what happens to your genome when a genomics company fails. The whole catastrophe originated in password reuse and a permissive sharing feature, not in any hack of 23andMe itself.

Why 23andMe still matters

23andMe is the lesson that you can be breached without being hacked. When your users reuse passwords, you do not enforce MFA, and you do not detect a login flood, attackers walk in through the front door using credentials you never held. And a feature that exposes data about other people means a few thousand accounts can leak millions of records. The data here, DNA and ancestry, is uniquely permanent and sensitive, and unlike a password it can never be reset. The defences: enforce MFA, which would have stopped the takeovers cold; detect and throttle credential stuffing and check passwords against known-breached corpora; design data-sharing features so one account cannot bulk-scrape millions of others (rate limits, minimised exposure); and minimise and protect uniquely sensitive data. The reused passwords that fuelled it trace back to mega-breaches like Yahoo.

How to fix it

  • Force password resets, enforce MFA, and invalidate sessions; a reset is the only way to evict reused-credential access.
  • Throttle and block the credential-stuffing traffic, and check user passwords against known-breached-password corpora.
  • Rate-limit and constrain the data-sharing features so a single account cannot scrape millions of others' records.

How to avoid it

  • Enforce MFA so a reused or stolen password is not enough on its own; it would have stopped this entirely.
  • Detect and throttle credential stuffing, and reject passwords known from prior breaches at sign-up and login.
  • Design social and data-sharing features so one compromised account cannot bulk-scrape data about other people.
  • Minimize and strongly protect uniquely sensitive data (genetic, biometric), and limit how much relatives' data any feature exposes.
  • Monitor for spikes in failed and unusual logins, the signature of an automated credential-stuffing run.

References

Related vulnerabilities

All OpSec →