Gartner projects that legal disputes involving technology companies will rise by 30% by 2028 due to violations of artificial intelligence (AI) regulations. The forecast reflects increasing concern among IT leaders about compliance risks as generative AI (GenAI) tools become more widely implemented.
The findings are based on a Gartner survey conducted between May and June 2025 among 360 IT leaders overseeing the deployment of GenAI tools. More than 70% of respondents identified regulatory compliance as one of their top three challenges in scaling GenAI productivity assistants. Only 23% said they were very confident in their organisation’s capacity to manage security and governance requirements during GenAI rollouts.
According to Lydia Clougherty Jones, Senior Director Analyst at Gartner, variations in global AI regulations have created inconsistent compliance obligations that complicate the alignment of AI investments with enterprise value and expose companies to additional liabilities.
The survey also found that geopolitical factors are influencing GenAI strategies, particularly outside the United States. Among non-US respondents, 57% said the geopolitical environment had at least a moderate impact on GenAI strategy and deployment, while 19% described the impact as significant. Despite this, nearly 60% reported that their organisations were either unable or unwilling to adopt non-US GenAI alternatives.
Gartner’s research further highlights the growing importance of AI sovereignty — the concept of national control over AI development, deployment, and governance. In a September 2025 webinar poll, 40% of 489 respondents said their organisations viewed AI sovereignty positively, while 36% expressed a neutral stance. Two-thirds indicated that their organisations were actively responding to sovereign AI strategies, and over half reported making structural or strategic changes as a result.
Gartner advises IT leaders to reinforce moderation mechanisms within GenAI systems in response to evolving legal and geopolitical pressures. Recommended actions include training models for self-correction, implementing rigorous use-case reviews, conducting extensive model testing with interdisciplinary teams, and introducing content moderation tools such as abuse reporting functions and AI warning labels.
