Organizations building trustworthy AI are 60% more likely to double the ROI of AI projects, highlighting the high cost of ignoring responsible practices
Carrie, North Carolina, September 30, 2025 /PRNewswire/ -SAS, SAS, a global leader in data and AI, has published a new study examining the use, impact and reliability of AI. IDC Data and AI Impact Report: Trust ImperativeCommissioned by SAS, we found that IT and business leaders report greater trust in generated AI than any other form of AI.
A global study examining AI use and adoption found that while organizations prioritizing trustworthy AI are 60% more likely to double the ROI of AI projects, only 40% invest in AI systems trustworthy through governance, explanation and ethical protection measures. Paradoxically, among those reporting the least investment in reliable AI systems, Genai (e.g. ChatGpt) was considered 200% more reliable than traditional AI (e.g. machine learning), despite being the most established, reliable, and explainable form of AI.
“Our research shows contradiction. Regardless of actual reliability or accuracy, the form of AI with human interaction and social familiarity seems to promote maximum trust.” “As AI providers, experts and individual users, we have to ask. Genai is trusted, but is it always reliable? And are leaders applying the guardrails and AI governance practices needed for this new technology?”
For more information, see more on LinkedIn Live Panel today. The Atlantic Ocean and Features IDC's SAS, Deloitte and guest speakers.
Visit our full research report: http://sas.com/ai-impact.
This study is based on a global survey of 2,375 respondents conducted in North America, Latin America, Europe, the Middle East, Africa and Asia-Pacific. Participants included a balanced combination of IT professionals and business line leaders, providing perspectives from both technology and business functions.
Emerging AI technology evokes most trust
Overall, this study found that more reliable AI deployments are emerging technologies like more established AI, such as Genai and Agent AI. Almost half of respondents (48%) reported “complete trust” on Genai, while the third said the same about Agent AI (33%). The least reliable form of AI is traditional AI, with fewer than one in five (18%) showing full confidence.
Even reporting high confidence in Genai and Agentic AI, survey respondents expressed concerns, including data privacy (62%), transparency and explanationability (57%), and ethical use (56%).
Meanwhile, even if the technology that performs most use cases is not yet fully realized, quantum AI is gaining confidence right away. Almost a third of global decision makers are familiar with Quantum AI, with 26% reporting full confidence in technology despite the fact that the actual application is still in its early stages.
AI Guardrails delays weaken AI shock… and ROI
This study showed a rapid increase in AI usage. In particular, it rapidly covers traditional AI in both vision and applications (81% vs. 66%). This raised new levels of risk and ethical concerns.
IDC researchers have identified inconsistencies in all regions about how reliable they trust AI and how reliable technology is. Research shows that nearly eight in 10 (78%) organizations claim they have full trust in AI, but only 40% have invested in making their systems obviously reliable through AI governance, explanation and ethical protection measures.
The study also showed that prioritization of reliable AI measurement implementation when operating AI projects is low. Of the top three organizational priorities of respondents, only 2% were selected to develop an AI governance framework, and less than 10% reported the development of responsible AI policy. However, robbing trustworthy AI measures may be preventing these organizations from becoming fully aware of AI investments in the future.
The researchers divided survey respondents into trusted AI leaders and trusted AI followers. Leaders seemed to invest most and earn rewards in practices, technology and governance frameworks to make AI systems trustworthy. These same trusted AI leaders were 1.6 times more likely to report more than twice as much ROI in their AI projects.
Strong Data Foundations and the Lack of Governance Stall AI
As AI systems become more autonomous and deeply integrated into critical processes, the fundamentals of data become more important. Data quality, diversity, and governance directly impact AI outcomes, making smart data strategies essential to realize benefits (such as ROI, increased productivity) and mitigate risk.
This study identified three major hurdles that hinder success in AI implementation. There is weak data infrastructure, poor governance, lack of AI skills. Almost half of organizations (49%) cite the basis of data that is not a non-centered or unoptimized cloud data environment as a major barrier. This biggest concern was followed by a lack of adequate data governance processes (44%) and a shortage of skilled professionals within the organization (41%).
Respondents reported issue No. 1 regarding the management of data used in AI implementations due to difficulty in accessing relevant data sources (58%). Other major concerns include data privacy and compliance issues (49%) and data quality (46%).
“For the benefit of society, businesses and employees, trust in AI is essential,” said Brian Harris, Chief Technology Officer at SAS. “To achieve this, the AI industry needs to increase the success rate of implementation, humans need to critically examine AI outcomes, and leadership strengthens the workforce with AI.”
About SAS
SAS is the global leader in data and AI. SAS software and industry-specific solutions allow organizations to transform their data into trustworthy decisions. SAS gives you the power to know®.
SAS Innovate 2026 – Unique experiences for business leaders, technology users, and SAS partners – will take place April 27-30, 2026 in Grapevine, Texas. Visit the SAS Innovate website for more information and saving dates!
SAS and all other SAS Institute Inc. products or service names are registered trademarks or trademarks of SAS Institute Inc. in the United States and other countries. ® Indicates US registration. Other brand names and product names are trademarks of their respective companies. Copyright©2025 SAS Institute Inc. All Rights Reserved.
Logo-https://mma.prnewswire.com/media/1250367/sas_v1_logo.jpg
