Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsAI should be judged not only by what its models can do, but by what happens to people when they use it. At a Salesforce Dreamforce side session in San Francisco, American Psychological Association CEO Dr. Arthur C. Evans Jr. argued that AI should serve human flourishing rather than become an end measured only by capability or speed. That shifts the design question from “Does the system work?” to “Does it help people function, make choices, build skills, and stay connected?”
What does it mean to flourish?
Evans’s framing is broader than feeling good. Flourishing includes functioning well and being connected to other people. That makes the effects of AI on confidence, critical thinking, judgment, decision-making, trust, relationships, autonomy, agency, and the full breadth of people’s work part of the design problem—not secondary considerations after technical performance.
As an Amazon Associate I earn from qualifying purchases.
In his March 2026 article for APA Monitor, Evans wrote, “Psychology is essential to getting it right.” The point is practical: understanding human behavior and experience is necessary to assess whether an AI system is serving people, not merely whether it completes a task.
Recommended Free Tools
Why technical success can still fail people
A system can operate exactly as intended and still be confusing or frustrating in use. Evans’s example is a door that is technically functional but gives people the wrong cues about whether to push or pull. The problem is not that the door cannot open; it is that its design does not fit how people interpret and use it.
#1 Best Overall
AI systems can create the same gap. A tool may return an answer quickly while weakening a person’s confidence in their own judgment, shifting responsibility in an unclear way, or changing how colleagues and customers relate to one another. Observing actual use—and listening to the people affected—reveals problems that technical tests alone may miss.
Questions to ask at each stage of AI adoption
For developers: what human effects belong in evaluation?
Alongside measures of technical performance, developers can ask whether a system supports or diminishes human abilities and skills. Evans put the opportunity this way: “We can be intentional about designing into the technology things that actually could help our abilities and our skills.” That requires evaluating how people use the system, not assuming that a successful output automatically benefits its user.
- Is the system helping people exercise critical thinking, or encouraging them to accept outputs without judgment?
- Does it strengthen confidence and decision-making, or make responsibility and accountability harder to understand?
- Does it support autonomy and agency, including the ability to question, decline, or override its suggestions?
For leaders and deployers: what changes for people and work?
Before introducing AI, examine the full job and the surrounding human system. Automating one task can alter how people coordinate, learn, make decisions, or understand their role. Ask, “What is the impact on the human systems?” and “What is the impact on the humans?” Include the people doing the work in identifying changes to autonomy, agency, relationships, and skill.
For customer-facing organizations: what happens to trust?
When AI mediates contact between an organization and its customers, consider what happens to the relationship as well as the transaction. Ask whether the interaction creates more trust or less trust, and whether people can tell when they are interacting with AI and how to reach a person when needed. Fast responses do not, by themselves, show that customer relationships are improving.
Rank #3
What APA’s 2026 psychologist survey says about chatbot use
The APA’s 2026 survey offers a concrete example of why human effects merit attention. It covered more than 1,200 licensed psychologists in the United States: 1,242 responded to invitations sent to more than 22,000 psychologists, with a reported completion rate of 6.3%. The online, probability-based random sample was fielded April 9–26, 2026. The percentages below describe psychologists’ reports about their patients—not the proportion of all people who use AI or chatbots.
| What psychologists reported | Survey finding |
|---|---|
| Patients discussed using AI for diagnosis, emotional support, or conversation | 77% of surveyed psychologists reported this among their patients |
| Patients used AI to self-diagnose | 39% of psychologists reported this |
| Patients used AI as an additional mental health professional | 35% of psychologists reported this |
| AI was used to assist treatment | 33% of psychologists reported this |
| AI was used for self-discipline, affirmations, or behavioral reminders | 34% of psychologists reported this |
| Patients used chatbots for fun, friendship, or intimate relationships | 33%, 22%, and 13% of psychologists, respectively, reported these uses |
| Among psychologists whose patients had ongoing chatbot conversations, patients talked about or appeared to feel supported or validated | 68% of those psychologists reported this |
| Among that same group, patients used chatbots to reinforce healthy coping skills | 41% of those psychologists reported this |
| Comfort with some patients engaging with chatbots | 54% of psychologists said they were comfortable with this |
| Concerns about certain patients using chatbots | 93% of psychologists had concerns |
| Observed chatbot dependency among patients engaging with chatbots | 36% of psychologists reported observing this |
| Observed distorted thinking or delusions related to a chatbot | 15% of psychologists reported this |
These are clinicians’ observations and judgments, not direct patient-prevalence estimates or proof that chatbots caused a particular outcome. The APA cautions that the survey cannot show the full scope of people using chatbots for mental-health advice or companionship. The mix of reported support and concern is a reason to ask more careful questions about how people are using these systems, not a basis for claiming that chatbots either improve or harm everyone’s wellbeing.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Keep mental-health guidance specific to mental health
The APA says AI is not a safe or effective substitute for a qualified mental-health provider. Its guidance for people navigating AI-generated advice also warns that AI can sound confident while being wrong, and recommends using it to complement a relationship with a licensed human professional. This is health-specific guidance; it should not be stretched into a blanket claim about every use of AI.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
A broader test than capability or speed
Evans’s argument is a design principle, not evidence that any particular AI intervention causes better flourishing, cognition, work, or relationships. Its value is in widening the test: alongside asking whether a system can perform, ask how it affects people’s functioning, skills, choices, trust, and connections in practice. Human outcomes are not an optional layer added after the technology is finished; they are part of what it means for the technology to work.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
