Source: DOJ is investigating whether the use of a predictive child welfare algorithm in PA's Allegheny County discriminates against parents with disabilities
For the two weeks that the Hackneys' baby girl lay in a Pittsburgh hospital bed weak from dehydration, her parents rarely left her side … Tweets: @apocalypticanow , @ap , and @rebeccakmccray Tweets: @apocalypticanow : “A recent analysis by ACLU researchers found that when Allegheny's algorithm flagged people who accessed county services for mental health and other behavioral health programs, that could add up to three points to a child's risk score, a significant increase on a scale of 20.” https://twitter.com/... @ap : Did an algorithm play a role in placing a baby in foster care? Two Pittsburgh parents wonder if an artificial intelligence tool singled them out because of their disabilities. The Justice Department is asking the same question, the @AP has learned. https://apnews.com/... Rebecca McCray / @rebeccakmccray : An algorithm used by a PA county child welfare agency is disproportionately flagging parents with disabilities and Black parents for investigation, predicting they're more likely to remove a child from the family https://apnews.com/...
Context & Ripple Effects
Allegheny County's risk-scoring tool has been under scrutiny for a year: experts warned in May 2022 that it could harden racial disparity in the child welfare system, and an ACLU analysis found that merely accessing county mental health or behavioral health services can add up to three points to a child's risk score — significant on a scale of 20.
What changed with this report is who is asking the questions: the Justice Department is now formally probing whether the algorithm discriminates against parents with disabilities, a shift from academic and advocacy criticism to federal civil-rights enforcement, centered on the Hackneys' case of a baby placed in foster care.
First-order effects
- The DOJ investigation puts Allegheny County under direct federal civil-rights examination over allegations that its tool disproportionately flags parents with disabilities and Black families for investigation — the county must now answer to an enforcer, not just critics.
- The ACLU's finding that behavioral-health service use itself raises risk scores becomes central evidence: the very act of seeking county help is penalized in the model.
Second-order effects
- Other jurisdictions running similar screeners — New York City's neighborhood-based abuse-risk tool has already drawn racial bias complaints — face pressure to audit their models before a federal probe arrives at their door.
- Vendors and counties deploying predictive tools in benefits and child welfare will likely see demand for disability-impact testing and service-use variables removed from scoring, following the pattern where states like Virginia, Georgia, and Tennessee dropped Family-Match after an investigation found it falling short.
Third-order effects
- If the DOJ finds discrimination, algorithmic screening in public services shifts from optional local experimentation to a regulated practice with federal liability attached — echoing how Arkansas's automated health assessments cutting disabled patients' home care and criminal-justice algorithms setting patrols and sentences have each triggered accountability fights.
- The structural question the probe forces: whether agencies can keep using proxies like service access and neighborhood in high-stakes decisions about families, or whether civil-rights law effectively bans such features from public-sector scoring models.
The trend: Predictive scoring in public services — child welfare, benefits, policing — is entering an era of federal civil-rights enforcement, where deployment without bias auditing increasingly carries legal consequences.