- Consequences in the past 12 months include operational disruption, financial loss, reputational damage and security breaches
- However, 92% say AI has delivered tangible benefits, while 87% expect to increase investment
- Rising costs and gaps in governance pose challenges for businesses looking to scale AI responsibly
October 22, 2026: Nearly seven in ten (69%) U.S. businesses have experienced an AI failure in the past 12 months, according to new research from BSI. The findings, based on a study of U.S. business leaders, including C-suite executives with insight into their organizations’ use of AI, shows that while the technology is delivering benefits, failures remain a reality for many.
For businesses that experienced failures, the effects were felt in several areas. Six in ten (60%) reported impacts on efficiency, such as downtime, while others cited financial loss (32%), regulatory or compliance issues (30%), safety failures (29%), security breaches (26%), and reputational damage (23%).
Despite these challenges, the business case for AI remains strong. More than nine in ten (92%) said their businesses' AI investments had delivered tangible benefits, including growth, innovation and efficiencies, while 84% said their AI projects and tools are delivering a return on investment.
The findings also suggest that while AI is delivering value for many organizations, managing its costs, establishing clear accountability and responding effectively when failures occur remains key to protecting employees, customers, and the overall business success.
Managing the cost of AI and protecting returns
The business case for AI comes with growing cost pressures. Seven in ten (71%) U.S. business leaders said the cost of AI had increased compared with 12 months ago, and 75% expect the cost of using AI to rise further.
At the same time, 87% expect their organization to increase its investment in AI. However, more than half (53%) said investment in AI had taken resources that could have been used on other projects, highlighting the choices businesses face as they expand their use of the technology.
For business leaders, the challenge is not simply whether to invest in AI, but how to make that investment sustainable. Clear governance, effective risk assessment, standards and guidance, appropriate staff training and processes for managing incidents can help organizations pursue innovation while preparing for the consequences when AI systems fail.
Building the foundations for responsible AI
The research points to opportunities for businesses to strengthen the processes that support AI use. Just 39% of businesses said their organization has an established AI governance program. While 56% of U.S. businesses reported having policies and processes for the safe and ethical use of AI, 66% said their organization has a standardized way to assess whether AI tools are acting as intended and delivering what is needed.
Meanwhile, 57% have a process for managing AI incidents and ensuring a timely response, leaving nearly half without a stated process for responding to timely incidents.
As businesses explore more autonomous AI systems, security and governance are also prominent concerns. Half (50%) cited security concerns about agentic AI, while 38% pointed to a lack of AI governance and 28% to the absence of national or global regulation. These concerns underline the need for organizations to consider how AI systems are monitored and managed as their use and investment continues to expand to new programs and areas of the business.
Staff training is another part of the picture. Although 78% were confident their organization is training staff on large language model outputs, such as ChatGPT and Copilot, less (only 55%) say their organization has a learning and development program to support successful AI training, pointing to the importance of translating confidence in AI into consistent skills and practices across the organization.
Susan Taylor Martin, Chief Executive, BSI, said: “As AI reshapes our businesses, economies and society, business leaders need to consider what happens when AI doesn’t work as intended. What’s clear is that AI failure is becoming a regular occurrence, bringing with it significant costs. We need to encourage leaders to do everything they can to prevent failures while being comfortable sharing what they have learnt when an issue arises.
“Realizing the benefits of AI requires taking responsibility for how it is used. Business has been here before, responding to growing cybersecurity risks by embedding safeguards proactively and rigorously investigating incidents. We must now put robust governance and guardrails in place to enable safe innovation that protects us all. Acting early can help reduce exposure to financial and legal liabilities, while also helping to build trust.”
Confidence in AI practices does not always translate into clear processes and consistent safeguards. Three quarters (75%) were confident their organization provides transparency over the data sources and algorithms used to train and deploy AI, while 50% said they know what sources of data their business uses. 72% were confident their organization implements safe and ethical AI policies, however, only 27% were confident their business is currently taking steps to align with global best practices and standards such as ISO/IEC 42001.
ENDS
Notes to editors
Alongside these U.S. findings, BSI has also explored the wider global picture, including the potential economic impact of AI failure and the importance of trust as organizations weigh the opportunities and risks of AI. For a broader perspective, read BSI’s global report, The Cost of AI Failure.
