Despite 84% of executives expressing confidence in AI-generated output without human review, just 11% believe their organisation’s data quality is sufficient for AI use, exposing a potentially dangerous gap between confidence, governance and reality.
Businesses may be placing greater trust in artificial intelligence (AI) than their underlying data and governance systems can justify, according to new global research revealing that AI-generated errors are already making their way to boards and external audiences. The Workiva 2026 Midyear Executive Benchmark Survey found that 84% of executives are at least somewhat confident in the accuracy of AI-generated output without human review.
Yet that confidence sits alongside a strikingly different reality. Just 11% believe their organisation’s data quality is sufficient for AI use, while one in four executives (26%) say internal audits have detected AI errors that reached external audiences or board members. The findings expose a potentially significant disconnect between executive confidence in AI and organisations’ ability to verify the information those systems produce.
As businesses accelerate AI adoption across finance, sustainability, risk and corporate reporting, the research raises a bigger responsible-business question. Are organisations adopting AI faster than their governance, data and human oversight can keep up?
EXECUTIVE AI CONFIDENCE COLLIDES WITH REALITY
The scale of executive confidence revealed by the research is notable. Despite growing concerns around AI hallucinations, inaccurate outputs and poor underlying data, more than eight in 10 executives surveyed said they had at least some confidence in AI output without human review.
At the same time:
- 26% said internal audits had identified AI errors that had already reached external audiences or board members.
- Just 11% believed their organisation’s data quality was sufficient for AI use.
- 27% said poor data quality had significantly blocked AI deployment in key workflows.
- 71% said poor data quality had at least moderately affected AI use in financial and sustainability reporting.
- 89% of institutional investors expressed concern about the accuracy of AI-generated content in corporate disclosures.
- The findings suggest the next phase of enterprise AI adoption may be less about whether organisations can deploy the technology and more about whether they can trust, verify and defend what it produces.
Barbara Larson, Chief Financial Officer at Workiva, warned that confidence alone is not enough. “Confidence in AI without control over data quality is a liability, not a strategy. CFOs need platforms that connect AI to trusted, auditable data so every output is one they can verify and every disclosure is one they can defend,” said Larson. “Getting this right is about more than avoiding errors. Business leaders can move faster and embed AI deeper into their operations when they trust what their systems produce. That’s a real competitive edge.”
AI ERRORS ARE BECOMING A BUSINESS RISK
The findings add to mounting evidence that AI risk is moving beyond hypothetical concerns and into everyday business operations. Earlier this month, Fair Play Talks reported that one in 10 publicly reported technology incidents now involves AI, with researchers warning that many organisations may not yet have the governance and resilience frameworks required to manage rapidly evolving AI-related risks.
There are also growing concerns about the quality of the work AI produces. Recent research found that two-thirds of employees spend up to six hours each week correcting poor-quality AI-generated work or “workslop”. This suggests that organisations still need robust quality assurance and human oversight before promised productivity gains can be fully realised. Separate research found that almost half of employees worry AI is making them less intelligent, adding to concerns about over-reliance on the technology and the potential erosion of critical thinking and human judgement.
The Workiva findings take that debate a significant step further. If inaccurate AI-generated information is reaching boards, investors or other external stakeholders, errors are no longer simply an employee productivity problem. They can become a corporate governance, reporting, reputational and accountability risk.
AI IS ALSO CHALLENGING EXISTING FINANCIAL CONTROLS
AI is not only creating risks through inaccurate output. It is also making some forms of manipulation easier. Research covered by Fair Play Talks in July found that 40% of US employees and 29% of UK workers had used AI to create or manipulate expense receipts.
Among US workers surveyed, 19% said they had used AI to fabricate a purchase entirely, while 15% had increased the value of an existing purchase. The findings raised questions about whether traditional financial controls are equipped to identify increasingly convincing AI-generated documentation. That challenge becomes even more significant when combined with workplace attitudes towards financial misconduct.
More recent research found that nearly nine in 10 finance professionals had ignored suspected workplace fraud. The Medius Financial Census 2026 found that 87% of finance professionals surveyed had ignored a small expense, reimbursement or claim they believed was fraudulent.
Together, these findings point to risks from two directions: technology can make manipulation easier while workplace cultures and controls may make questionable behaviour easier to overlook. For finance leaders in particular, responsible AI therefore cannot be separated from internal controls, organisational ethics, data integrity and accountability.
POOR DATA COULD UNDERMINE AI TRUST
The Workiva research also highlights an issue that can easily become obscured by enthusiasm surrounding generative AI: AI output is only as trustworthy as the information and governance supporting it. Just 11% of executives surveyed believe their data quality is sufficient for AI use.
That matters particularly when AI is being incorporated into financial reporting, sustainability disclosures and other information on which boards, investors and regulators may rely. Almost three-quarters (71%) said poor data quality had at least moderately affected their organisation’s use of AI in financial and sustainability reporting.
Institutional investors appear acutely aware of the potential consequences, with 89% concerned about AI accuracy in corporate disclosures. If businesses cannot demonstrate where information originated, how an AI system reached a conclusion and whether its output was appropriately checked, questions about AI accuracy can quickly become questions about corporate accountability and stakeholder trust.
HUMAN OVERSIGHT STILL MATTERS
As AI systems and autonomous agents become more sophisticated, the Workiva research suggests businesses recognise that supporting infrastructure will remain essential.
Among executives surveyed:
- 49% said organisations will continue to require systems of record such as general ledgers.
- 45% said software enabling traceability and audit will remain necessary.
- 55% said platforms will be needed to manage AI agents and automated workflows.
Jason Darby, Chief Financial Officer at Amalgamated Bank, said specialised systems and human judgement remain important when AI is used in financial reporting. “Generic AI isn’t enough for financial reporting. Investors, regulators, and boards expect answers they can trust,” said Darby. “The real advantage comes from specialised AI built on governed, auditable data and paired with human judgment, giving organisations the confidence to verify what AI produces and stand behind the decisions and disclosures that follow.”
The emphasis on human judgement is significant. The challenge for organisations is not necessarily to insert a person into every AI-assisted task, but to determine where human oversight is essential, who remains accountable for the final decision and how AI-generated information can be verified.
WORKERS ARE ALREADY MOVING FASTER THAN GOVERNANCE
The Workiva findings also connect with another emerging workplace challenge: employees themselves are adopting AI faster than many organisations can establish rules around its use.
As Fair Play Talks recently reported, workers are embracing AI faster than employers can govern it. Across 72 organisations analysed by Prodoscore, employees were actively using almost 50 different AI applications, with consumer platforms including ChatGPT, Gemini and Claude among the most widely used.
The findings highlighted the rise of “shadow AI” – employees using AI tools outside company-managed technology environments – creating potential risks around confidential information, intellectual property, compliance and the quality of AI-generated work. Employees who feel uncomfortable disclosing their AI use further complicate that governance challenge.
Recent research revealed that half of Gen Z workers feel guilty about using AI and more than four in 10 are hiding their AI use from employers. That creates a difficult situation for organisations. Businesses cannot effectively govern AI if leaders do not know which tools employees are using, employees do not understand what constitutes acceptable use, or workplace cultures encourage people to conceal their reliance on the technology.
AI TRANSPARENCY REQUIRES PSYCHOLOGICAL SAFETY
Technology policies alone are unlikely to solve that problem. As Fair Play Talks has previously explored, AI transparency increasingly depends on psychological safety in the workplace. Employees need to feel able to disclose AI use, question an AI-generated answer and report mistakes without fearing embarrassment, punishment or damage to their professional reputation.
Otherwise, organisations risk creating exactly the wrong incentives: employees may conceal their use of AI or quietly correct problems themselves rather than raising issues that could expose wider weaknesses. For employers, this makes workplace culture part of responsible AI governance. Policies can specify which tools employees should use, but organisations also need environments in which people feel able to challenge AI rather than automatically trust it.
LEADERSHIP CONFIDENCE IS NOT THE SAME AS AI READINESS
Perhaps the most important lesson from the Workiva findings is that confidence should not be mistaken for capability or control. That distinction is emerging repeatedly across workplace AI research. A recent global study found that 70% of workers believe they are ready to work with AI, while only 39% of business leaders believe employees are prepared.
Meanwhile, businesses may still be underestimating the people challenges created by AI transformation, with organisations frequently concentrating on technology deployment while giving insufficient attention to workforce readiness, skills and organisational change.
Taken together, the findings suggest an organisation can be technologically ambitious without necessarily being organisationally ready.
THE COST OF GETTING AI GOVERNANCE WRONG
There is also a commercial dimension to the governance debate. Businesses are under growing pressure to demonstrate that substantial investments in AI are producing measurable results. Earlier research found that seven in 10 companies could slash AI budgets if returns continue to disappoint. Nearly 70% of executives surveyed said they were prepared to reduce AI spending if business goals were not achieved, while 73% reported that at least some AI investments had failed to meet expectations during the previous 12 months.
The research also highlighted the growing burden associated with AI oversight, governance and implementation. That matters because inaccurate outputs, poor-quality data, duplicated work, verification requirements, compliance failures and time spent correcting AI-generated mistakes all potentially undermine the productivity gains organisations hope AI will deliver. Responsible AI governance should therefore not be viewed simply as an additional compliance cost.
Good governance may ultimately be part of the business case for AI itself. If organisations cannot trust AI outputs, demonstrate their provenance or prevent employees from spending substantial amounts of time checking and correcting them, promised returns on AI investment become considerably harder to realise.
WHAT EMPLOYERS & BUSINESS LEADERS SHOULD DO NEXT
The findings suggest responsible AI governance can no longer sit solely with IT. As AI becomes embedded across finance, HR, sustainability, risk, communications and other business functions, organisations need governance structures capable of connecting technology decisions with data quality, accountability, employee behaviour, legal obligations and human judgement.
Strengthen data governance before scaling AI
Organisations should understand the quality, provenance and reliability of the data feeding AI systems before deploying those systems into high-stakes workflows. Poor-quality data should be treated as an AI governance risk rather than simply a technical inconvenience.
Define where human review is mandatory
Not every AI-assisted task carries the same level of risk. Financial disclosures, regulatory reporting, employment decisions, legal work and other high-impact activities may require significantly stronger human oversight. Businesses should identify those activities explicitly rather than leaving individual employees to decide when AI-generated output requires checking.
Make AI outputs auditable
Organisations should be able to establish where important information originated, how it was processed and who reviewed or approved the final output. Traceability becomes increasingly important as autonomous AI agents undertake more complex tasks across organisations.
Strengthen financial controls for the AI era
Existing controls may not have been designed for a world in which employees can generate convincing documents and receipts using AI within seconds. The rise of AI-generated expense fraud demonstrates why organisations may need to reassess fraud detection, verification processes and employee guidance as generative AI becomes more sophisticated.
Create clear accountability
AI may generate an answer, recommendation or document, but responsibility cannot simply be transferred to an algorithm. Organisations need clarity about who remains accountable when AI informs a decision, report or disclosure – particularly when that information reaches customers, regulators, investors or boards.
Understand how employees actually use AI
Formal enterprise systems tell only part of the story. The growth of shadow AI means organisations need better visibility into how employees are using consumer AI tools while creating a culture in which appropriate use can be discussed openly.
Create a culture where people challenge AI
Employees should feel able to question an AI-generated answer rather than assume that because a system produced it, it must be correct. As Fair Play Talks has highlighted, psychological safety is becoming an important component of AI transparency. Organisations should encourage employees to raise errors, uncertainty and concerns early rather than rewarding unquestioning reliance on AI-generated output.
Build AI literacy across leadership
Executives and boards need enough understanding of AI’s capabilities and limitations to challenge outputs rather than defer to them. The Workiva findings demonstrate why this matters: high confidence without adequate data, controls or verification can itself become a governance risk.
Measure the hidden costs of AI
Organisations evaluating AI ROI should look beyond software licences and headline productivity measures. Time spent checking outputs, correcting mistakes, managing risk, investigating incidents and strengthening governance should also form part of the calculation. The question should not simply be “How much AI are we using?” but “Is AI helping us produce better, safer and more reliable outcomes?”
WHY THIS MATTERS FOR RESPONSIBLE BUSINESS
AI has moved rapidly from experimentation into everyday business operations. The governance surrounding it now needs to catch up. The Workiva findings expose an uncomfortable contradiction. Executives are highly confident in AI-generated output, yet relatively few believe the data underpinning those systems is actually ready for AI – and some organisations are already discovering errors after they have reached boards or external audiences.
That should matter to more than technology teams. When AI-generated information influences financial reporting, sustainability disclosures, investment decisions or strategic choices, inaccurate outputs can affect employees, investors, customers, regulators and wider stakeholder trust.
And this is increasingly part of a broader pattern. Workers are adopting AI faster than employers can govern it. Organisations are encountering new AI-related operational risks. Employees are spending hours correcting poor-quality AI-generated work. AI is being used to circumvent traditional financial controls. And businesses are under mounting pressure to prove that their AI investments actually deliver value.
Responsible AI therefore requires more than adopting the latest technology or publishing an AI policy. It requires good data, transparency, traceability, appropriate human judgement, clear accountability, psychological safety and leaders willing to question what AI tells them.
The biggest risk may no longer be that businesses are moving too slowly on AI. It may be that their confidence in it is moving faster than their ability to verify whether it is right.
Download the full Workiva report here.
RELATED ARTICLES
One in 10 reported technology incidents now involves AI, according to a new report.
Two-thirds of employees say they spend up to six hours a week fixing low-quality, AI-generated work, according to study.
A growing number of employees fear that AI is making them less intelligent, according to a new survey.
New research reveals growing financial pressures are driving workers in the US and UK to manipulate expenses as employers face a new AI-powered fraud challenge.
Nearly night in 10 finance professionals admit they have ignored a claim they believed was fraudulent.
Employees are embracing AI faster than many organisations can govern it, according to new research.
Half of Gen Z workers feel guilty using AI and more than four in 10 are using it without their employer’s knowledge, despite businesses increasingly demanding AI skills from new recruits.
AI transparency policies will only succeed when employees feel psychologically safe to disclose how they use AI at work, argues Dr Gleb Tsipursky.
Global study finds companies accelerating AI adoption despite weak leadership readiness, trust gaps and rising pressure to prove ROI from workplace AI investments.
New global research exposes a growing gap between AI ambition and organisational readiness.
Global survey shows that nearly 70% of companies are prepared to scale back AI budgets this year if projects fail to deliver measurable business results.
Sign up for our newsletter
