“Psychological safety is a must in this new age of AI,” says Rafi Torfur, executive vice president and chief technology officer at Infosys. “Tech itself is evolving so fast.
To assess how psychological safety affects success with enterprise-level AI, MIT Technology Review Insights surveyed 500 business leaders. The findings show high self-reported levels of psychological safety, but also suggest that fear still has a step. Along with the story, industry experts highlight one reason for the disconnect between rhetoric and reality: While organizations may publicly promote a safe message to experiment with, deep cultural flaws can counter that intent.
Building psychological safety requires an integrated, systems-level approach, and human resources (HR) alone cannot make such a change. Instead, businesses should embed psychological safety deeply into their collaboration processes.
Key findings for this report include:
- Companies with experiment-friendly cultures have more success with AI projects. The majority of executives surveyed (83%) believe a company culture that prioritizes psychological safety significantly improves the success of AI initiatives. All four leaders agree that organizations that promote such safety are more successful in adopting AI, and 84% observe links between psychological safety and tangible AI outcomes.
- Psychological barriers are proving to be more barriers to enterprise AI adoption than technical challenges. Encouragingly, nearly three-quarters (73%) of respondents indicated that they feel safe to provide honest feedback and express opinions freely in their workplace. Still, a significant share (22%) admits that they are hesitant to lead an AI project because they might be blamed if something goes wrong.
- Achieving psychological safety is a dynamic goal for many organizations. Less than half of leaders (39%) rate their organization’s current level of psychological safety as “very high.” Another 48% report it to a “moderate” degree. This may mean that some enterprises are pursuing AI adoption on cultural grounds that are not yet fully established.
This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by the editorial staff of MIT Technology Review. It was researched, designed and conducted by human writers, editors, analysts, and writers. This includes writing the survey and collecting data for the survey. The AI tools that have been used were limited to secondary production processes that passed full human review.