62% of companies reported high returns from trustworthy AI — Cyprus Mail
Cypriot outlet Cyprus Mail reported on a study by SAS and IDC, according to which organizations with trustworthy artificial intelligence practices were 15 times more likely to report significant or high returns on investment in AI projects. Such results were reported by 62% of organizations investing in relevant measures, compared with 4% of organizations without such practices.
Governance and data quality
The authors of the second Data and AI Impact Report, titled The New Economics of Trust, found that organizations with the strongest governance, data quality, and audit practices reported at least twice the returns from AI implementation compared with other survey participants. At the same time, this group represented a relatively small part of the market.
Organizations with the highest scores under the trustworthy AI criteria also reported 1.85 times greater improvements across 13 business outcomes, including revenue growth, cost reduction, and customer experience. Only 17.5% of enterprises had fully optimized data infrastructure, which researchers considered sufficiently mature for the needs of agentic AI.
Organizations with an optimized data foundation were four times more likely to expect high returns from AI projects. They were also six times more likely to require data quality controls and decision explainability designed to strengthen trust in systems.
More current news is available on the UA.News Telegram channel Telegram.
Employees often override AI recommendations
According to the report, 97.2% of users at least sometimes override AI-generated recommendations. The most common reason was the system's inability to explain the logic behind its decision, regardless of whether employees considered the result correct.
Trust in AI declined as system autonomy increased: it stood at 76% for generative AI and 66% for agentic AI. SAS Chief Technology Officer Bryan Harris noted that the error rate of current agents in complex tasks may exceed 25%, which is unacceptable for high-risk decision-making.
The study is based on a global survey of 2,699 people who are knowledgeable about or influence their organizations' data and AI initiatives. It covered 28 countries and four industries: banking, insurance, life sciences, and the public sector.
