Automated decision-making technology (ADMT) is a regulatory term for technology that processes personal information and uses computation to replace or substantially replace human decision-making. It became a defined term of art in U.S. state privacy law through regulations issued under the California Consumer Privacy Act (CCPA), and is used more broadly to describe the class of automated and AI-assisted systems that make or materially inform consequential decisions about individuals.
Definition under the CCPA regulations
The California Privacy Protection Agency (CPPA) defines ADMT as any technology that processes personal information and uses computation to replace or substantially replace human decision-making. The definition expressly includes profiling and excludes routine tasks — such as web hosting, spam filtering, and simple data organization — unless those tools are used to make a covered decision (Source: mayerbrown.com; ogletree.com).
The regulations attach obligations when ADMT is used for "significant decisions" affecting individuals — areas such as employment, housing, credit, education, and access to essential goods and services. Covered businesses must provide a pre-use notice describing the use of ADMT, offer a right to opt out of that use, and provide a right to access information about how the technology was applied to the individual (Source: skadden.com; paulhastings.com).
Rulemaking and effective dates
The CPPA completed the ADMT rulemaking on September 22, 2025, as part of a package that also covered cybersecurity audits and risk assessments. The regulations began taking effect on January 1, 2026, with the ADMT-specific compliance obligations phased in so that businesses using ADMT for significant decisions must meet the notice and opt-out requirements by January 1, 2027 (Source: cppa.ca.gov; paulhastings.com).
Relationship to other regimes
ADMT obligations overlap conceptually with algorithmic accountability requirements and with the meaningful human review standards appearing in other AI-governance instruments, and they form one strand of the state-level activity discussed under AI federalism. Comparable profiling and automated-decision provisions appear in other state privacy laws and in the EU's data-protection framework, making ADMT a recurring reference point in debates over how AI-assisted decisions about individuals should be governed (Source: skadden.com).
Relationships
- depends-on: AI and Privacy — ADMT is a privacy-law construct.
- related: Algorithmic Accountability and Bias Audits — overlapping accountability goals.
- related: Meaningful Human Review — human-oversight standard for automated decisions.
- related: AI Federalism — part of state-level AI regulation.
- related: California CCPA/CPRA Regulations — Automated Decision-Making Technology (ADMT) — regulations defining ADMT and its obligations.