EY: The Problems & Solutions of CEOs Misjudging AI Concerns

According to EY, there is a gap between what senior executives believe and actual public concerns about AI, with many executives being misinformed about the issues that matter most.

When EY compared responses from executives with a separate survey of more than 15,000 consumers across 15 countries, the gap was bigger than expected. 

On almost every measure of responsible AI – from data accuracy to privacy protection – ordinary people were roughly twice as worried as CEOs.

The disconnect is academic as well as threatening to undermine the multi-billion pound AI boom as companies pour resources into technologies that consumers might reject.

In response, EY proposes an award winning framework to guide enterprises to AI success.

A key problem: Mature AI adopters showing overconfidence

For some time, companies have been racing to integrate large language models (LLM) into everything from customer service to financial planning, but EY’s research suggests they’re building on shaky foundations of public trust from the start.

For instance, regarding the issue of job losses, both bosses and consumers rank this relatively low on their worry list, despite years of media coverage about AI replacing workers. 

Instead, the public frets about fake news generated by AI, the potential for manipulation and whether vulnerable people might be exploited.

“AI implementation is unlike the deployment of prior technologies,” says Cathy Cobey, EY’s Global Head of Responsible AI.

“It’s not a ‘one-and-done’ exercise but a journey, where your AI governance and controls need to keep pace with investments in AI functionality.”

According to EY, the problem gets worse among companies that consider themselves AI veterans. 

Among firms claiming to have “fully integrated” AI across their operations, 71% of executives believe they understand consumer sentiment.

That compares with just 51% at companies still getting to grips with the technology.

Additionally, the people in EY’s research that are newer to the field of AI are more in sync with public opinion.

Whereas executives at companies still implementing AI show genuine concern about privacy, security and reliability – worries that mirror those of consumers.

EY says that once companies have deployed AI systems and worked through the initial teething problems, their leaders seem to assume the public has moved on too.

The company’s consumer research also reveals massive gaps between people’s theoretical willingness to use AI and their actual behaviour, particularly in sensitive areas like banking and healthcare where trust matters most.

The study also exposes another truth that roughly one in three executives claims their company has “fully integrated and scaled AI solutions”– showing wishful thinking rather than reality.

This means that true AI integration involves far more than plugging in the latest chatbot.

It requires overhauling business processes, retraining staff and building robust data governance systems.

EY’s nine principles to close AI governance gaps

In response to these findings, EY has developed a nine-point responsible AI framework that addresses the specific areas where companies are falling short.

The principles span:

• Accountability
• Data protection
• Reliability
• Security
• Transparency
• Explainability
• Fairness
• Compliance
• Sustainability

The framework tackles consumer concerns head-on.

While data protection ensures AI systems maintain confidentiality of personal information whilst reflecting ethical norms, transparency requires appropriate disclosure about AI system purpose and design so users can understand and evaluate outputs. 

Additionally, explainability demands that decision criteria can be reasonably understood and challenged by human operators.

EY’s research says companies currently maintain strong controls for just three out of these nine areas on average. 

The gap is particularly pronounced in fairness – assessing impacts on all stakeholders to promote inclusive outcomes – and sustainability, which embeds consideration of physical, social, economic and planetary impacts throughout the AI lifecycle.

Next wave of AI amplifying governance headaches

According to EY, the challenges are about to get harder. 

Companies are gearing up for agentic AI – systems that can make decisions and take actions without human oversight. 

However, half the executives surveyed admit their current risk management approaches won’t cope with these more powerful systems. 

EY says that on average, companies have strong controls in place for just three out of nine key areas of responsible AI.

Furthermore, more than 51% say it’s already difficult to create proper oversight for today’s AI tools, let alone future ones. 

Yet many companies planning to deploy advanced AI in the next year haven’t even familiarised themselves with the associated risks.

Cathy says: “Maintaining trust and confidence in AI will require continuous education of consumers and senior leadership, including the board, on the risks associated with AI technologies and how the organisation has responded with effective governance and controls.”

CEOs grasping consumer concerns better than colleagues

Despite the stark gap between CEO’s understanding of AI concerns, there’s one bright spot in EY’s data: CEOs seem more plugged in to public sentiment than their fellow board members. 

CEOs show the best alignment with consumer concerns about responsible AI and are least likely to claim their companies have bulletproof controls.

Furthermore, half of CEOs say they take primary responsibility for AI strategy – more than any other role except Chief Technology and Information Officers. 

CEOs also spend more time with customers than most other senior executives, giving them better insight into public opinion.

The pattern suggests a communication problem within companies. If CEOs understand the issues but other executives don’t, it points to a failure to cascade concerns down through the organisation.

EY’s three-step solution: Listen, act, communicate

To address these challenges, EY proposes a three-pronged approach that goes further than traditional risk management.

Step 1
The first step involves exposing the entire C-suite to customer voices. 

This means getting back-office executives like CTOs and CIOs into customer-facing situations. 

EY suggests putting these leaders in focus groups, exposing them to customer surveys and in healthcare companies, mandating that all senior executives spend time interacting with patients in clinical settings.

Step 2
The second phase requires integrating responsible AI throughout the development process. 

Companies should build on existing practices like human-centric design and A/B testing to create “human-centric responsible AI design” as an integral element of innovation.

This goes beyond regulatory compliance to address specific customer concerns raised by AI applications.

Step 3
The final element focuses on communication. 

EY argues that responsible AI represents a competitive differentiator rather than just a compliance burden.

Companies that transparently explain their safeguards and governance processes can stand out from competitors who remain silent about their AI practices.

The trust gap creating a competitive opportunity

The research suggests that companies see responsible AI as a compliance burden, but EY’s data suggests it could be a competitive weapon.

Consumers have genuine worries about AI safety, but most companies aren’t explaining how they address these concerns. 

That creates an opportunity for firms willing to be transparent about their safeguards and governance processes.

The alternative is a race to the bottom where consumers lump all companies together and assume the worst about their AI practices.

Cathy sees the potential upside: “By taking the lead on responsible AI and making it a centrepiece of your brand and messaging, you can stand out in the eyes of your current and potential customers and position yourself ahead of the competition.”