- A recent global study reveals a striking paradox: trust in generative AI is accelerating rapidly due to its humanlike responsiveness, even as the technology exhibits lower reliability and reduced explainability compared to traditional machine learning systems.
- Agentic AI adoption is already widespread (over 50%), yet its actual performance remains low, creating a crucial need for robust AI governance to secure financial returns and mitigate project risks.
(hightechPRIME.com) — A recent global study conducted by IDC reveals a striking paradox in the adoption of generative artificial intelligence (genAI): trust in genAI is accelerating rapidly, driven primarily by its humanlike responsiveness, even though it exhibits lower reliability and reduced explainability compared to traditional machine learning systems. The SAS-sponsored survey, which gathered insights from 2,375 IT and business leaders across multiple sectors, found that organizations with minimal investment in AI governance perceived genAI as up to 200% more trustworthy than conventional AI technologies.
This surge in perceived trust stands in stark contrast to persistent implementation challenges. Research from Carnegie Mellon University (CMU) and the Massachusetts Institute of Technology (MIT) indicates that a significant majority of AI pilot initiatives fail to meet expectations. CMU’s experimental simulation, “TheAgentCompany,” tested leading AI agents on routine office tasks and revealed that even the most advanced systems struggled with basic operations such as closing pop-up windows or interpreting standard file formats. On average, these agents successfully completed only 25% of assigned tasks, underscoring the gap between perceived capability and actual performance.
Despite these limitations, the adoption of agentic AI—autonomous systems designed to operate in dynamic environments—has already reached over 50% among surveyed organizations. However, IDC warns that without robust frameworks for governance, ethics, and transparency, companies risk diminished returns on investment and stagnation in AI-driven innovation. Supporting this caution, Gartner forecasts that by 2027, approximately 40% of agentic AI projects will be discontinued due to escalating costs, unclear value propositions, and insufficient risk mitigation strategies.
IDC emphasizes that trust in AI is not merely an ethical imperative but a financial necessity. Organizations that establish dedicated governance teams and invest in responsible AI platforms are 60% more likely to achieve a twofold increase in project ROI. The report also highlights the emergence of quantum AI—a fusion of quantum computing and artificial intelligence—which is generating considerable interest across industries such as climate science, financial services, and logistics. Although still in its experimental phase, quantum AI is viewed as a promising frontier for solving complex problems previously deemed computationally infeasible.
As generative and agentic AI systems become increasingly embedded within core enterprise workflows, IDC asserts that the true differentiator will be integration. This entails harmonizing structured and unstructured data, enforcing governance protocols, and embedding explainability into automated decision-making processes. Without these foundational elements, the transformative potential of AI may remain unrealized, leaving organizations vulnerable to inefficiencies and reputational risks.
