Global Trust in Generative AI Surges Despite Lack of Safety Measures
Trust in Generative AI: A Global Surge Amidst Concerns
Overview of the Research
SAS, a leader in data and AI, recently released findings from their IDC Data and AI Impact Report titled "The Trust Imperative." This research reveals a significant increase in trust towards generative AI (GenAI) technologies among global business leaders and IT professionals.
Over 2,000 respondents participated in this comprehensive study, which highlighted the paradoxical trend: while trust in generative AI is soaring, investment in responsible AI practices is lacking. Despite 60% of organizations prioritizing trustworthy AI systems being twice as likely to see a substantial return on investment (ROI) from AI projects, only 40% are currently investing in the necessary governance frameworks, ethical guidelines, and transparency measures.
The Discrepancy Between Trust and Responsibility
Interestingly, GenAI like ChatGPT is perceived as 200% more trustworthy than traditional AI solutions, despite the latter being long-established, reliable, and interpretable. Kathy Lange, Research Director at IDC, points out the contradiction in perceptions of these AI technologies. The characteristics of human-like interaction and social familiarity associated with generative AI make it appear more trustworthy, even when its reliability doesn't match that perception.
Insights from Global Decision Makers
The study categorizes participants into trustworthy AI leaders—those who heavily invest in governance and ethical practices—and followers, who lag behind. Leaders report a 1.6 times higher likelihood of doubling their ROI on AI projects compared to their less responsible counterparts. Yet, concerning areas such as data privacy (62%), transparency (57%), and ethical usage (56%), are fears among those reporting high trust in generative AI. Particularly, the study found that while almost 80% of organizations claim to trust AI, only 40% have taken measures to ensure that their systems are trustworthy.
The Impact of Emerging Technologies
Emerging AI technologies, particularly GenAI, are leading in trust perception, with 48% of respondents expressing "complete trust" in these systems compared to only 18% for traditional AI. The rapid emergence of GenAI has led to increased visibility and application, raising ethical concerns and risks not previously encountered. Decision-makers express a strong desire for robust governance frameworks to prevent potential pitfalls associated with increased reliance on AI.
Despite the growing confidence in technologies like quantum AI, which 26% reported complete trust in despite its nascent stage, the need remains for comprehensive safeguards to bridge the gap between perceived trust and actual reliability.
Barriers to Trustworthy AI Investment
The research identifies several critical barriers for organizations attempting to build trust in AI systems. These include poor data infrastructure, ineffective governance processes, and a shortage of skilled personnel. Nearly half of the surveyed participants identified unoptimized data environments as major impediments to successful AI implementation.
Another significant concern was the lack of clear AI governance policies, with only 2% prioritizing governance frameworks in their operational strategies. Without adequate attention to ethical considerations and transparency in AI development, companies risk stalling growth and diminishing the potential ROI from their AI investments.
Recommendations for Moving Forward
To enhance trust in AI, SAS emphasizes the essentiality of developing clear success metrics and appropriate review processes for AI implementations. According to Bryan Harris, CTO of SAS, it's critical for organizations to empower personnel with AI tools, promote critical evaluations of AI results, and amplify the success of AI strategies.
Going forward, SAS will lead discussions on these findings and the future of AI through platforms like LinkedIn Live, aiming to change how organizations perceive and trust emerging technologies.