Early stage, but a clear acceleration
Nearly 1,000 safety professionals completed a survey conducted by the American Society of Safety Professionals (ASSP) between July 6 and August 2, 2026. ASSP is a professional and trade association based in Park Ridge, Illinois (United States) that also offers its own training and certification programs. The findings appear in the report “Trust, Ethics and Governance: How Safety Leaders Are Thinking About and Using AI.” According to ASSP, 86 percent of respondents are in an exploratory or transformational phase of an AI strategy for workplace safety, and two-thirds of them only started in the past two years. Organizations are mainly using AI for reporting, training, incident analysis, predictive analytics, automated compliance and real-time monitoring, trade publication ISHN reports. In ASSP’s view, this shift means AI is moving from an experimental tool to a fixed part of day-to-day safety work, though the organization stresses that most participants are still very much testing and scaling up.
Culture and data quality as decisive factors
Respondents cited low or early investment in resources (74 percent), data quality (62 percent), lack of ownership or priority (59 percent), distrust or insufficient confidence (55 percent), and concerns about privacy, security or bias (also 55 percent) as the main barriers. According to ASSP, organizational culture, the importance assigned to AI, and fewer ethical objections are strong predictors of AI adoption; the report also links training to greater knowledge and confidence. ASSP president Monique Parker emphasizes, in the article, the importance of human oversight of AI applications above all.
Caveat: this is not independent research. It was carried out by ASSP and funded by its own ASSP Foundation; the organization itself offers training and certification programs in this area. The figures are therefore self-reported by its own membership, the recruitment method and response rate are not explained, and the claimed “acceleration” is not backed by historical comparison data. The results mainly reveal how this particular group of professionals views AI, not necessarily how the sector as a whole does.
Context
Independent research tempers this enthusiastic picture. The National Institute for Occupational Safety and Health (NIOSH) in Washington (United States), part of the federal public health agency CDC, conducts its own research into the risks of AI in the workplace through its Future of Work Initiative, and points to uncertainties around oversight, liability and the quality of training data. An Australian study by Cebulla, Szpak and Knight (2023) in the International Journal of Workplace Health Management, based on interviews with experts, organizations and workplace inspectors, describes psychosocial risks, work stress, privacy risks and risks to business continuity when AI is introduced. According to the authors, these risks can recur throughout the entire life cycle of an AI system, and they call for a systematic risk assessment at every stage of implementation. This independent perspective raises concerns about governance and data reliability that the ASSP report itself does not resolve. For EHS departments considering AI tools, that is at least as relevant as the adoption figures.
Source: ISHN, “New Research Finds AI Adoption Accelerating in Workplace Safety” (15-10-2026).
Also consulted: NIOSH Future of Work Initiative (CDC) and academic research in the International Journal of Workplace Health Management on WHS risks of AI in the workplace.
