What it changes
Who can pull it
What it looks like institutionally
Most professional AI use now runs ahead of workplace guidance: in the national social-work survey, 63.5% of respondents already use AI tools in their role while 26.6% report no training for those tools, 53.4% want training, and 66.8% want ethical-use guidelines. The gap is not enthusiasm — it is literacy: knowing what the tool cannot do, when its output needs checking, and where its output may not go.
AI literacy as a governance pattern has two halves. The trust half calibrates reliance: staff who know a system's failure modes defer to it less blindly, which is why the field literature proposes AI literacy as a core competency even for non-users. The boundary half makes data rules taught rather than merely posted — which case details may enter which tools, and why pasting client information into an unvetted external assistant is a disclosure, not a shortcut.
Literacy dampens ungoverned egress; it does not close the pathway. Closing it structurally — allow-lists, authorization, egress controls — is connection authorization's job, and the two patterns pair naturally: the rule people understand plus the pathway that enforces it. Pair the trust half with real verification skill (see Verification training): knowing the tool's limits tells you when to check; training tells you how.
Ledgered PAN-run results used above
The same national survey 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]
AI literacy — the knowledge and skills required to understand, use, and critically evaluate AI systems — has been proposed as a core competency for social work, relevant even to practitioners who never directly use AI tools.[†]
In a national survey of 1,179 U.S.-based social workers conducted from October 2025 to February 2026 by the University of Texas at Austin in collaboration with NASW, 63.5% of respondents reported using AI tools or technologies in their current role.[2]
Addresses: Uncalibrated trust in AI output · Ungoverned use ahead of guidance · Case detail pasted into unvetted tools. Test a version of this lever in the PAN Lab.