AI Adoption Is Now a People Risk: What the People Risk 2026 Report Means for UAE Companies

Sep 3

9 min read

Short answer: The Marsh People Risk 2026 Global Infographic shows that AI adoption is not only a technology issue. It is connected to cyber literacy, skills shortages, leadership capability, data and IP handling, regulatory change, and employee mindset. For UAE companies, the practical lesson is to treat AI implementation as a workforce risk programme first, then as a technology rollout.

What the Report Actually Says

The Survey Base

The People Risk 2026 Global Infographic surveyed 4,517 respondents, including 2,258 Risk professionals and 2,259 HR professionals. The survey covered 26 markets across Asia, Europe, Latin America and the Caribbean, the Middle East and Africa, North America, the Pacific, and the United Kingdom. The UAE sample included 103 respondents.

The Risk Method

The report assessed 25 key people risks across five pillars. HR and Risk professionals assessed the likelihood and severity of those risks for their organization over the next one to two years. The product of those scores was used to produce a Risk Rating Score.

The five people risk pillars are:

Pillar	Relevance to AI implementation
Technological change and disruption	AI changes workflows, skills, risk exposure, and operating models.
Talent, leadership, and workforce practices	AI adoption depends on skills, leadership behaviour, and workforce readiness.
Governance, compliance, and financial	AI introduces data, IP, accountability, and regulatory questions.
Protection, environment, and sustainability	AI use increases the importance of cyber literacy and security-minded behaviour.
Health, well-being, and safety	AI-driven change may interact with mental health, job displacement, and adaptation pressure.

The Global People Risks Most Relevant to AI Adoption

In the combined global ranking for HR and Risk professionals, the top 10 risks were:

Rank	Global people risk	Why it matters for AI implementation
1	Inadequate cyber threat literacy	Employees need to understand how AI can expose data, systems, and confidential information.
2	Labor shortages	Skills scarcity affects the ability to build, operate, and govern AI-enabled workflows.
3	Technology skills shortages	AI projects require technical, operational, and managerial capability.
4	Employee financial insecurity	Workforce pressure can affect behaviour, productivity, and security-minded actions.
5	Increasing health and benefit costs	Benefits and health pressures compete with transformation budgets and management attention.
6	Mindset barriers to AI adoption	AI value depends on whether employees and managers understand, trust, and use AI correctly.
7	Mishandling of data and IP	AI tools increase the importance of clear data boundaries and IP controls.
8	Changing regulatory environment	AI adoption must be documented and governed as regulation evolves.
9	Mental health deterioration	Organisational change and work pressure may affect employee readiness and resilience.
10	Uncompetitive talent strategies	AI capability becomes part of talent attraction, retention, and development.

The key point is not that AI is one isolated risk. The report shows AI adoption sitting inside a broader system of people risks: skills, leadership, cyber behaviour, data handling, regulation, and workforce resilience.

What the Report Says About AI Specifically

The infographic states that artificial intelligence is reshaping expectations of work and accelerating organizational change.

It also states that many employers focus on AI dangers such as data loss and hallucinations, while a greater threat is failing to convert AI investment into meaningful productivity, innovation, and performance gains.

The report identifies two specific concerns related to a lack of AI mindset:

Finding	Meaning for AI implementation
40% of HR and Risk professionals are concerned about investment in AI without appropriate employee training and upskilling.	AI tool access without workforce capability creates implementation risk.
38% are concerned about lack of AI knowledge in HR limiting transformation of people practices.	HR capability matters because AI changes roles, skills, performance, and workforce planning.

The infographic also groups AI-related risks into workforce and organizational risks.

Risk group	Risks listed in the infographic
Workforce risks	Security/cyber incidents, misinformation and disinformation, regulatory changes, lack of AI mindset, systems and safety failures
Organizational risks	Discrimination and bias, mental health deterioration, obsolete skills, job displacement

What This Means for UAE Companies

(let’s look at it from our perspective)

For UAE companies, the report should be read as a warning against treating AI implementation as a narrow software procurement decision.

The UAE is included in the survey sample, but the public infographic does not publish a UAE-only ranking. Therefore, the correct interpretation is:

The global findings are relevant to UAE companies because UAE organizations face the same structural AI adoption questions: cyber literacy, skills, data handling, leadership capability, employee mindset, and regulatory change.

This is especially relevant for companies that operate internationally, serve regulated sectors, or work with EU, UK, or multinational clients.

AI Adoption Is a Workforce Capability Problem

The report supports the view that AI adoption is a workforce capability issue, not only a technology issue.

Why?

Because the risks most connected to AI adoption are not only model risks. They are people and operating-model risks:

AI adoption challenge	Connected people risk from the report
Employees using AI without understanding data boundaries	Inadequate cyber threat literacy; mishandling of data and IP
Teams unable to redesign workflows around AI	Technology skills shortages; mindset barriers to AI adoption
Managers unable to supervise AI-supported work	Inadequate leadership skills
HR unable to adapt roles and training	Lack of AI knowledge in HR
AI tools deployed without measurable productivity gain	Failure to convert AI investment into performance gains
Employees anxious about automation	Obsolete skills, job displacement, mental health deterioration

The report does not say that every AI programme will fail without these controls. It does show that the underlying people risks are material enough to be ranked and measured by HR and Risk professionals globally.

What Changes for Employees When AI Enters the Organisation

AI implementation changes work in practical ways. Employees may move from producing every output manually to supervising, validating, and improving AI-assisted outputs.

Before AI	With AI-enabled work
Employee writes from scratch.	Employee reviews and improves AI-generated drafts.
Employee searches manually.	Employee checks retrieved and summarised information.
Employee creates reports manually.	Employee validates assumptions, sources, and business meaning.
Employee follows a fixed process.	Employee decides when AI use is appropriate and when human judgment is required.

This shift requires more than tool access. It requires:

AI literacy,
data classification awareness,
source verification,
prompt discipline,
escalation rules,
human review standards.

The report supports this direction by showing that 40% of HR and Risk professionals are concerned about AI investment without appropriate employee training and upskilling.

Why AI Literacy Is Different From Tool Training

Tool training answers: “How do I use this product?”

AI literacy answers:

What data can I use?
What data should never enter an AI tool?
When can I rely on an AI output?
When must I verify sources?
When must I involve a human expert?
What does the model not know?
What should I do if the output is plausible but wrong?

A one-hour product demo is not enough for employees who handle confidential business, customer, financial, legal, or operational information.

What Changes for Managers

The report states that leadership skills and digital fluency are increasingly crucial for effective risk management. It also states that inadequate supervisory and leadership skills trigger or worsen more risks than any other people risk, based on global data.

In AI implementation, this means managers become the practical control layer between strategy and daily work.

Managers need to decide:

Manager responsibility	AI implementation question
Workflow redesign	Which tasks should AI support, automate, or leave human-led?
Quality control	What does a good AI-assisted output look like?
Risk escalation	When should employees stop and ask for review?
Employee confidence	Do people understand what AI is for and what it is not for?
Adoption discipline	Are teams using approved workflows or informal tools?

Without trained managers, AI adoption becomes inconsistent. One team may use AI well, another may avoid it, and another may use unapproved tools without clear rules.

What Changes for HR

The report says 38% of HR and Risk professionals are concerned that lack of AI knowledge in HR limits the transformation of people practices.

[Diuna interpretation] HR should therefore not be limited to sending training invitations after the tool has already been rolled out.

HR should help define:

HR responsibility	AI adoption implication
Skills mapping	Identify who needs basic, intermediate, or advanced AI capability.
Role design	Update responsibilities where AI changes workflows.
Training	Move from generic awareness to role-specific AI literacy.
Performance management	Reward correct AI use, not only faster output.
Employee relations	Address anxiety, fairness, workload, transparency, and confidence.
Leadership development	Train managers to supervise AI-supported work.

AI implementation is a workforce operating-model change. HR should be part of that design.

What Changes for Risk and Cybersecurity

Inadequate cyber threat literacy is the number 1 global people risk in the combined HR and Risk ranking.

Mishandling of data and IP is number 7 globally.

These findings are directly relevant to AI adoption because AI tools create new ways for employees to misuse, expose, or mishandle information.

Examples include:

AI-related behaviour	Risk
Pasting confidential data into unapproved AI tools	Confidentiality and privacy exposure
Uploading contracts for summarisation	Legal and IP exposure
Using AI browser extensions without review	Third-party data exposure
Using AI-generated code without review	Security vulnerabilities
Using AI-generated research without source checks	Misinformation and decision risk
Connecting AI agents to business systems too early	Excessive agency and operational risk

These examples are not listed directly in the infographic. They are practical examples of how the report’s cyber, data/IP, and AI mindset risks may appear during implementation.

What UAE Companies Should Do First

UAE companies should start with a controlled AI adoption model, not a broad AI rollout.

The first executive question should be:

Which AI use is officially allowed, for whom, under which data rules, and with which review process?

That answer should be short, operational, and understandable to managers and employees.

First 30-Day Operating Model

Step	Action	Primary owner
1	Identify current AI usage and informal workarounds	IT + department heads
2	Classify AI use cases by risk	Risk + operations
3	Approve 2–3 low-risk workflows	Executive sponsor
4	Define data rules for AI use	Legal + security
5	Train managers first	HR + AI owner
6	Launch role-based AI literacy	HR + department leads
7	Create an escalation channel	AI owner + security
8	Review adoption, quality, and incidents monthly	Executive sponsor

The objective is not to slow AI down. The objective is to make safe AI usage easier than unsafe AI usage.

Recommended AI Use Case Sequence

Start with workflows that create visible value but limited downside.

Phase	Use cases	Why they come first
Phase 1	Internal writing, meeting summaries, internal policy explanations	Low technical complexity, visible productivity gain
Phase 2	Knowledge-base Q&A, sales enablement, support agent assist	Requires better data governance and review
Phase 3	AI-assisted reporting, workflow automation, internal copilots	Requires integrations, access control, and monitoring
Phase 4	Customer-facing AI, decision support, autonomous agents	Requires stronger governance, testing, accountability, and incident response

Starting with autonomous agents or high-impact customer workflows before employees understand AI basics creates avoidable implementation risk.

The Right AI Adoption KPIs

The report states that organizations that successfully manage and mitigate people risks reported positive outcomes including:

Positive outcome	Share of respondents
Increased workforce productivity and efficiency	40%
Faster progress on strategic initiatives, e.g. AI adoption and sustainability	36%
Greater resilience during external shocks	32%
Tangible financial savings or improved profitability	29%

AI adoption metrics should therefore measure more than tool access.

Better implementation KPIs include:

KPI	Why it matters
Employees trained by role	Measures readiness, not only access.
Approved AI use cases live	Shows controlled adoption.
AI outputs reviewed before external use	Measures quality discipline.
Reported AI incidents	Shows visibility and trust.
Manual effort reduced in defined workflows	Measures actual workflow impact.
Shadow AI cases replaced by approved workflows	Measures risk reduction.
Manager confidence in AI-supported work	Measures leadership readiness.
Employee confidence in AI boundaries	Measures clarity and adoption discipline.

What Companies Often Overlook

Based on the report’s focus on cyber literacy, AI mindset, skills shortages, leadership, data/IP, and regulatory change, companies should be careful not to overlook the following issues.

1. Training Comes Too Late

If employees receive access before rules, the company creates avoidable ambiguity.

2. Everyone Gets the Same Training

Finance, HR, sales, legal, support, and engineering do not face the same AI risks. They need different examples and approval paths.

3. Managers Are Not Prepared

Managers translate AI policy into daily behaviour. If they are not trained, adoption becomes inconsistent.

4. Resistance Is Treated as Irrational

Mindset barriers may reflect real uncertainty about work quality, role changes, workload, or job displacement.

5. AI Is Added on Top of Broken Workflows

AI creates value when the workflow changes. If the process remains unchanged and AI is simply added on top, the result may be more noise, more review work, and unclear accountability.

Practical Framework for AI Implementation

Use five workstreams.

Workstream	Purpose	Minimum deliverable
Governance	Decide what AI use is allowed	AI usage rules and risk tiers
Workforce	Prepare employees and managers	Role-based AI literacy programme
Security	Protect data, IP, and systems	Approved tools, access rules, incident path
Workflow	Redesign work around AI	2–3 measurable use cases
Assurance	Monitor value and risk	Monthly adoption, quality, and incident review

This framework converts the report’s people risk findings into an implementation model.

About the author: Diuna Technologies — AI strategy, software engineering, and secure AI implementation team. Diuna supports companies with AI readiness assessment, AI strategy crafting, AI MVP delivery, and production AI implementation for international and regulated environments.

Scope and Source Note

This article is based on Marsh’s People Risk 2026 Global Infographic: The Human Edge — Transforming Risk into Strategic Advantage. The public infographic confirms a UAE sample of 103 respondents, but it does not publish a UAE-specific risk ranking table. UAE-specific implications and implementation steps are presented from Diuna's perspective;