What it is
A worker handed a new system learns to use it by using it, and nobody teaches them what its output gets wrong. The survey cited below describes that gap between exposure to AI and preparation for it. Use outruns training.
What it pushes on in the Lab
In the Lab, this pressure lowers the verification skill of the people using the system. It raises operator deference drift. It also raises how much of the system's failed output people adopt, because the newest staff lean on the tool most. It thins skill rather than adding cases, so it adds no work of its own.
Who feels it
The newest workers feel it as confidence they have not earned. The clients feel it when that confidence meets an error the worker was never taught to see.
What answers it
An answer has to strengthen the judgment people bring to the output, or put fewer unchecked answers in front of them. Each lever below does at least one of these.
Levers in the Lab that push the other way on something this pressure pushes on:
- Verify outputPattern:Put a verifier on the agent
- Review the riskiest firstPattern:Risk-tiered oversight
- Pause AI on alarmsPattern:Deployment circuit-breaker
- Escalate checksPattern:State-feedback vigilance
- Keep skills sharpPattern:Deskilling-arrest mandate
- Understand the systemPattern:Understand the system
- Incident loop
- Train the staff
The list leaves out levers the Lab has retired, levers it keeps as counter-examples, and any lever no network offers.
Where it starts switched on
Networks in the Lab that start with this pressure switched on:12
- CA-CDS Child Abuse Alerting
- CDTFA Axyom Assist
- Colorado Family Safety and Risk Assessments
- Credit Acceptance's net-collections Score, inside CAPS
- CrimSAFE criminal-record tenant screening
- Google's child-safety detection and account enforcement
- IDx-DR Autonomous Screening
- NJ AI Assistant
Every network that starts with it switched on
The evidence behind its effects
The Lab cites these claims from the evidence registry for this pressure's effects.
The 2025–2026 University of Texas at Austin / NASW national survey of U.S. social workers describes a gap between AI exposure and AI preparedness: 26.6% of respondents cited lack of training or understanding of AI technology as a challenge, 53.4% said training on AI tools and effective use would help, and clear guidelines on the ethical use of AI were the most-endorsed need (66.8%).[2]
Borah, E., Meyerhoff, J., Al-Turk, A., Gower, K., & Mastryukova, A. (2026, June). Use of artificial intelligence in social work practice: Findings and recommendations from a national survey. Moritz Center for Societal Impact, Steve Hicks School of Social Work, The University of Texas at Austin. https://moritzcenter.utexas.edu/wp-content/uploads/2026/06/Moritz-Center-AI-SW-Survey-Report.pdf
https://moritzcenter.utexas.edu/wp-content/uploads/2026/06/Moritz-Center-AI-SW-Survey-Report.pdf
Appears in: Evidence correction (2026)
Topics: human-ai-interaction, social-work
Pinazo-Hernandis, S., & Carcavilla-Gonzalez, N. (2026). Are future social workers ready for AI? Fears, barriers, and learning needs in higher education. Social Work Education. https://doi.org/10.1080/02615479.2026.2631708
doi.org/10.1080/02615479.2026.2631708
Appears in: Evidence reverification (2026)
Topics: human-ai-interaction, social-work