Why responsible AI is critical for trust and bottom-line performance

The high cost of ignoring AI risks

Responsible AI adoption has benefits but neglecting it can come at a steep cost. Almost every company in the global survey (99%) reported financial losses from AI-related risks with 64% experiencing losses in excess of US$1 million. On average, the financial loss to organisations that have experienced risks is conservatively estimated at US$4.4 million. That amounts to an estimated total loss of US$4.3 billion for the 975 respondents to the global survey.

The most common risks organisations reported are non-compliance with AI regulations (57%), negative impacts on sustainability goals (55%), and bias in outputs (53%). Issues such as explainability, legal liability, and reputational damage are less prominent but are expected to grow as AI becomes more visible and deployed at scale.

Irish respondents share similar concerns. Key concerns are in relation to AI usage were the lack of explainability in AI decision making and bias or discrimination in AI outputs – each cited by 63% of Irish respondents. 57% expressed fears in relation to a failure to disclose the purpose, design and impact of AI systems to stakeholders or end users was a matter of concern. The same number were worried about non-compliance with AI regulations. Cybersecurity vulnerabilities introduced by AI systems were cited by 54% of respondents with the same number pointing to misuse of AI by employees or customers as one of their main causes for concern.

Interestingly, high energy and resource use impacting sustainability commitments and goals was called out by 49% of respondents.

Irish organisations are taking steps to address the risks associated with AI usage. 

Lack of knowledge leaves companies exposed: Despite the mainly positive findings of our survey, it’s clear that many C-suite leaders don’t have sufficient understanding of the controls required to mitigate AI risks. While many organisations have controls in place, knowledge gaps persist. When asked to match the appropriate controls against five AI-related risks, only 12% of the respondents to the global survey got them all right. This lack of knowledge has consequences. There is a direct correlation between the scale of the loss incurred due to AI risks and a lack of appropriate controls. This highlights a clear need for targeted upskilling of the C-suite and not just those employees who will be using AI.

Agentic AI and citizen development: the new frontier

The AI risk environment is evolving rapidly, with the advent of agentic AI presenting new governance challenges. These systems operate continuously, adapt quickly, and require limited human intervention, making traditional oversight models insufficient. Most organisations are already implementing governance frameworks to address risks. Governance measures being implemented include continuous monitoring and incident escalation processes for unexpected agentic behaviours.

Irish organisations have a broadly positive view of agentic AI and citizen development with a majority of respondents saying their organisations are encouraging or at least permitting employees to develop AI agents. More than one third (34%) of respondents say their organisation actively encourages it while 14% permit it under specific conditions or use cases. For a further 17%, policies vary by department. Only 34% of respondents said their organisations prohibit its use.

Very importantly, the overwhelming majority (96%) of those organisations that encourage or permit agentic AI development have responsible AI guidance in place. In 39% of cases this is takes the form of formal, organisation-wide policies or frameworks aligned to responsible AI principles while the remainder provide informal guidance or suggested practices which are not standardised across the organisation.

Irish organisations also maintain a high level of vigilance in relation to agentic AI use with 57% of respondents saying a central inventory is kept of all agentic AI use cases and activities. In 40% of cases, the respondents say their organisation has moderate visibility with awareness of some, but not all, agentic AI use cases.

A critical aspect of agentic AI that is being overlooked by many organisations is the advent of the hybrid AI-human workforce. This will present challenges for workforce models but only a third (32%) of global respondents say their HR team is developing strategies for managing such environments. At the same time, this does indicate that companies are beginning to think strategically about the longer-term possibilities and workforce implications of the technology.

The emergence of agentic AI with all of its implications for risk and workforce organisation highlights the need for responsible AI to evolve in step with technology and workplace behaviours. Defined frameworks, proactive oversight and strong leadership understanding are critical if organisations want to realise the tangible bottom-line benefits of responsible AI.

Actions for Irish business leaders

Here are three actions Irish business leaders can take to strengthen their organisations’ AI governance and controls, and boost business performance:

Summary

The EY Global Responsible AI Pulse survey highlights that organisations adopting responsible AI achieve measurable gains in efficiency, innovation, and customer responsiveness, while those neglecting governance face considerable losses. Strong governance built on clear principles, robust controls, and effective oversight drives tangible gains in revenue, cost efficiency, and employee satisfaction. Irish organisations show early progress with budgets, training, and monitoring in place. Emerging agentic AI introduces new risks and that makes leadership upskilling and proactive policies critical for sustainable competitive advantage.