AI Has Entered the I‑O Mainstream—Now Comes the Hard Part
May 15, 2026 in AI, Change, Transparency, & Communication, Informed Decision-Making, Innovative Capabilities, Leadership & Influence, Strategic Planning, Technology & Tools, Workforce Planning
By Julie Jasewicz
This year’s Society for Industrial and Organizational Psychology (SIOP) conference made one thing unmistakably clear: artificial intelligence (AI) is no longer a fringe experiment in the field. In less than two years, AI has moved from an experiment to commonplace across nearly every I‑O discipline. The train has left the station and yet many organizations are still standing on the platform, unsure how to board responsibly.
Conversations at SIOP reflected both excitement and unease. There was significant discussion about ethical and environmental concerns tied to AI, paired with an uncomfortable truth: there is no single governing body or silver‑bullet solution coming to resolve these risks for us. Instead, the responsibility is landing squarely on individual organizations to understand what technologies they are using, how those tools were built, and what unintended harm they may create.
The future of AI in work is not inevitable. It is a set of choices, and those choices sit with industry leaders.
How AI Is Being Used Across I‑O Psychology
AI has rapidly expanded across traditional I‑O domains, often accelerating work that once took months to complete into tasks accomplished in hours. But speed does not equal rigor, and most SIOP presenters stressed that expert oversight remains essential.
Psychometrics and Scale Development
AI is increasingly used to:
- Generate initial item pools based on construct definitions
- Paraphrase items to improve reading level accessibility
- Flag potential subgroup bias or redundancy during early review
However, experts warned that AI frequently produces plausible, but psychometrically weak items. Human judgment is still required to ensure construct fidelity, validity, and fairness at every stage, aligning with SIOP’s published guidance on AI‑based assessments[1].
Training and Learning Design
Practitioners are leveraging AI to:
- Create just‑in‑time learning content tailored to role or skill gaps
- Generate scenarios and practice exercises at scale
- Summarize learner feedback and evaluation data
Used well, AI can expand learning access. Used poorly, it can flatten nuance and embed inaccuracies that learners may not know to question.
Workflow Augmentation
AI is now embedded in day‑to‑day I‑O work, to support:
- Drafting project plans, communications, and reports
- Summarizing interview notes or qualitative data
- Assisting with proposal development and internal documentation
This shift has pushed AI from “tool” to “teammate,” fundamentally changing accountability and decision ownership.
Data Analysis and Insights
AI supports:
- Rapid exploratory analyses and pattern detection
- Narrative summaries of statistical results
- Visualization suggestions for non‑technical stakeholders
Multiple presenters emphasized that AI often produces analytical “slop,” including outputs that look confident, but are directionally incorrect, necessitating an expert who is equipped to detect these errors.
Competency Modeling and Job Analysis
AI is being used to:
- Analyze job descriptions and occupational databases at scale
- Draft competency language aligned to strategic goals
- Identify skill adjacencies and emerging role requirements
AI can accelerate synthesis, but it cannot replace context, stakeholder input, or validation with incumbents.
Assessment Development
One of the most consistent threads at SIOP was assessment integrity in the age of AI. Practitioners demonstrated how AI is already being applied across the five phases of assessment development:
- Behavioral definition drafting
- Item generation
- Readability and bias review
- Content validity evaluation
- Pseudo-factor analysis and refinement
At the same time, concern is growing about AI‑assisted cheating, particularly in unproctored or remote assessments. Strategies discussed to mitigate risk included:
- Designing items that require contextual reasoning rather than recall
- Limiting exposure time and randomizing item presentation
- Pairing AI detection tools with human review, rather than relying on detection alone
The consensus: compliance and integrity must be designed into assessments from the start.
The Human Cost of “Move Fast” AI Adoption
AI has gone from experiment to enterprise expectation in under two years. The technology may be ready, but the infrastructure to support integration into workflows is not.
Research shows that 95% of AI pilots fail, overwhelmingly due to human and organizational factors such as insufficient training, unclear governance, and lack of workflow redesign, not because the models themselves don’t work.[2]
Meanwhile, the workforce impact is uneven:
- Entry‑level employees are bearing the brunt of early AI disruption through task automation and augmentation
- Workers who adapt effectively are already seeing substantial benefits, with AI‑skilled employees earning wage premiums of up to 56% according to PwC’s 2025 Global AI Jobs Barometer[3]
This widening gap underscores a critical ethical question: who is responsible for adaptation?
Research consistently shows that AI adoption succeeds when organizations are explicit about expectations, invest in training, and actively support employees through the transition. Companies that pair AI tools with clear guidance, leadership engagement, and role‑relevant training see significantly higher adoption, greater employee confidence, and stronger returns, while organizations that treat AI as a self‑service experiment struggle to move beyond pilots and informal use[4].
What Employees Are Actually Experiencing
Data shared at SIOP paints a complex picture of employee attitudes:
- 60% of employees trust AI outputs without verifying accuracy, increasing organizational risk[5].
- More than 50% of employees report that their organization has not clearly communicated how AI will impact their role, contributing to anxiety and uncertainty[6].
- 46% believe time saved using AI belongs to them, not the organization, highlighting emerging tensions around productivity, ownership, and expectations[7].
Trust, transparency, and fairness are becoming leadership issues not technical ones.
The Takeaway: Ethical AI Is a Leadership Choice
What emerged most clearly at SIOP is this: the future of AI at work is not pre‑written. It depends on how leaders choose to deploy AI and whether they invest in people at the same pace they invest in tools.
Responsible AI use in I‑O psychology means:
- Treating AI adoption as a change management initiative, not a software rollout
- Providing structured training so expertise, not blind trust, guides use
- Designing governance, validation, and accountability into systems from day one
- Recognizing that ethics, accuracy, and sustainability are ongoing commitments
AI can absolutely amplify good I‑O practice. But without thoughtful leadership, it can just as easily undermine trust, widen inequities, and erode the very rigor our field is known for.
The train has left the station. Where it goes, and who it leaves behind, is still very much up to us.
Julie Jasewicz joined FMP in May 2023 as a Human Capital Intern and works on a variety of projects ranging from training and development to strategic communications. She graduated from George Mason University’s IO psychology master’s program and is originally from the Adirondack mountain region in New York. Julie is passionate about cooking, travel and is a loving cat mom to her kitten