Aging-services professionals increasingly use artificial intelligence to draft case notes, summarize intake calls, schedule visits, prepare outreach, and organize benefits information. These tools can reduce administrative burden. They can also turn a weak assumption about an older adult into a polished recommendation that travels through an organization without enough human review.
The American Society on Aging unites and equips professionals working to improve aging, with a strong emphasis on equity, anti-ageism, and healthy longevity. That mission calls for a practical standard: before an AI-assisted output reaches an older adult, caregiver, clinician, case manager, or public agency, someone should complete an AI dignity check.
Why Accuracy Alone Falls Short
A response can contain accurate facts and still undermine dignity. An AI system may describe a person mainly through deficits, treat family support as automatic, assume low digital ability, or recommend the most restrictive option because it appears administratively simple. It may confuse chronological age with capacity, risk, preference, or willingness to learn.
An intake summary may omit an older adult’s own goals. A care-plan draft may elevate a caregiver’s preferences over the client’s. A benefits explanation may use language that discourages questions.
These errors can surface in familiar workflows. An intake summary may omit an older adult’s own goals. A care-plan draft may elevate a caregiver’s preferences over the client’s. A benefits explanation may use language that discourages questions. A scheduling tool may repeatedly offer digital-only options to someone who needs phone or in-person access.
ASA’s strategic principles treat older people as experts, agents for change, and storytellers in their own lives. AI-assisted work should preserve that agency rather than quietly converting professional convenience into a client constraint.
Use a Five-Part Dignity Check
First, identify the consequential decision. Separate low-stakes assistance from outputs that shape eligibility, services, safety, housing, healthcare, finances, or autonomy. The higher the consequence, the stronger the review.
Second, inspect the source material. Confirm that the system received current, relevant, and authorized information. Do not place protected health information, financial records, case histories, or confidential family details into an unapproved tool. A fluent answer cannot repair incomplete or improperly shared data.
Third, test for ageist assumptions. Ask whether the output equates age with frailty, dependence, resistance, cognitive decline, or technological inability. Check whether it treats family involvement as universally available or desirable. Review the language for paternalism, unnecessary urgency, and recommendations that narrow choice without evidence.
Fourth, restore the person’s voice. Compare the draft with the older adult’s stated goals, cultural context, communication preferences, risk tolerance, and definition of a good outcome. In benefits counseling, this may mean presenting options and tradeoffs rather than selecting one. In care coordination, it may mean documenting what the client declined and why. In outreach, it may mean offering multiple ways to respond.
Fifth, assign a named human owner. The NIST AI Risk Management Framework emphasizes governing, mapping, measuring, and managing AI risks. In an aging-services workflow, that means a qualified person approves the output, can explain the reasoning, corrects errors, and remains accountable for the decision.
Make Disclosure Safe
Many staff members already experiment with AI privately. They may hide that use because colleagues could view it as lazy, risky, or unprofessional. Hidden use prevents supervisors from seeing prompts, data choices, corrections, and failure patterns. It also keeps useful practices trapped at the individual level.
Leaders should create a simple disclosure norm: staff can state where AI assisted, what sources controlled the result, what they verified, and who approved the final output. This approach gives cautious employees a safe boundary and gives enthusiastic users a reason to slow down before sharing a polished mistake.
Invite skeptics into the review process. Staff members who notice loss of nuance, inaccessible language, cultural assumptions, or threats to client autonomy can help design stronger prompts and escalation rules. Their concerns supply quality control rather than resistance to progress.
AI can help aging-services professionals spend less time formatting information and more time listening, explaining, and coordinating.
Start With One Bounded Workflow
Choose one contained task, such as converting intake notes into a structured summary, drafting a follow-up message after benefits counseling, or preparing alternative versions of a caregiver resource. Establish the dignity check before the pilot begins.
Measure more than time saved. Track corrections, omitted preferences, inappropriate assumptions, client questions, staff confidence, and whether the workflow expands or narrows meaningful choice. Compare the AI-assisted process with the existing one.
AI can help aging-services professionals spend less time formatting information and more time listening, explaining, and coordinating. That benefit depends on keeping the older adult’s goals in charge. A dignity check turns that principle into a repeatable workflow.
Gleb Tsipursky, a behavioral scientist called the “Office Whisperer” by The New York Times, helps tech-forward leaders stop overpaying for AI while boosting engagement and innovation. He serves as the CEO of the AI consultancy Disaster Avoidance Experts, and wrote eight books, including “The Psychology of AI Adoption at Work: From Resistance to Results” (Georgetown University Press, 2026), from which this article was adapted.
Photo credit: Shutterstock/Bixstock
