The American Psychological Association released a detailed playbook on Monday for companies introducing artificial intelligence into the workplace, urging employers to treat trust, oversight and job design as core requirements rather than afterthoughts. The guidance, titled "Deploying AI into Workplaces: What Leaders Need to Know," lays out seven objectives for organizations and comes with survey data suggesting most workers have no idea what their employer's AI rules actually are.

The document arrives as AI adoption accelerates across white-collar industries and as employees report rising anxiety about what the technology means for their jobs. For companies weighing their own rollout, the APA's message is blunt: buying software and mandating its use is not a deployment strategy. Readers looking for more context on the broader shift can follow our ongoing AI developments coverage, which tracks policy, research and workplace adoption.

What the Survey Found

The guidance is anchored in APA's 2026 Work in America survey, conducted online by the Harris Poll among 2,009 employed adults in the United States between May 21 and June 1, 2026.

The headline number: fewer than 40 percent of workers for whom the question applied said they knew whether their employer had an official policy on acceptable AI use. In other words, most employees are working alongside AI tools without knowing what their organization permits, prohibits or expects.

Anxiety is climbing alongside adoption. The survey found 48 percent of workers worried AI could make some or all of their job duties obsolete — the highest level recorded since the question was introduced in 2023. Some 37 percent worried they would fall behind colleagues who use AI, up from 30 percent in 2025. And 64 percent believed technological change benefits employers more than workers, compared with 52 percent a year earlier.

"Purchasing an AI tool and telling people to use it without further direction is not a sound strategy," said Dennis P. Stolle, APA's executive lead psychologist for applied psychology. Employers, he argued, should tell workers where tools are reliable and where they are not, ensure oversight is meaningful, and rethink how jobs themselves are designed.

The Seven Objectives

The guide is explicit about its scope: it is written for organizations that have already decided to adopt AI, not for those weighing whether to. Its checklist covers introducing, using and evaluating AI tools, with examples of effective implementation and red flags.

The seven objectives span the life of a deployment:

  • Job design. Tasks and entire roles may need redesign, with attention to mental workload, attention, awareness of the wider work situation, hidden labor and incentives. The guide flags the amount and distribution of verification work in AI-assisted roles as an emerging evidence question.
  • Expertise. Organizations should preserve opportunities to develop deep knowledge, share it and learn from experience — and watch for skill decay and deskilling when people are pushed out of the loop.
  • Trust calibration. Workers need support in developing and adjusting their trust in systems over time, rather than being expected to trust or distrust a tool on day one.
  • Human oversight. Oversight must be meaningful rather than a formality of keeping a person in the loop.
  • Clear responsibility. Employers should clarify where accountability sits in mixed human-AI work, and match safeguards to the stakes of potential errors.
  • Fair process. Deployments should involve consultation with the people who will use a system and those affected by its outputs, with voice, transparency, consistency and the ability to correct outcomes.
  • Psychological boundaries. The final objective covers limits on monitoring, privacy, autonomy and dignity, and treats the validity of AI-enabled inferences about workers' internal states as an open evidence question.

What It Is — and What It Isn't

The APA is careful to position the document. It is not a compliance standard, not an APA professional standard, and not an evaluation of any product. Its evidence framework explicitly distinguishes decades of established research — on trust in machines, error detection, attention, job performance and organizational change — from newer, still-provisional findings about current AI systems.

The advisory panel that informed the guidance drew on a wide set of subfields, including human-technology interaction, clinical, social, legal, cognitive and industrial-organizational psychology.

How Employers Should Use It

The APA recommends running the checklist before a significant deployment and then revisiting it on a regular schedule — and specifically after a new system is introduced, a model changes, a major upgrade ships or the organization restructures. Organizations are encouraged to record important decisions, who owns them, and when they will be reviewed.

Which objectives matter most will vary by the work, the goals and the risks involved, the guide notes — but the underlying premise applies everywhere. AI deployment is not a procurement event with an end date; it is a continuing test of whether workers can judge a system's reliability, exercise real oversight, keep building expertise and participate in decisions as both the technology and their jobs change.

With most employees still unaware of whether AI rules exist at their workplace, the APA's message to employers is that closing that gap is the first, cheapest and most obvious place to start.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →