Business leaders across major global economies are increasingly concerned about the rapid rise of misinformation, cyber threats and the potential negative impacts of artificial intelligence, according to new findings from the World Economic Forum (WEF).

The WEF Executive Opinion Survey 2025, based on responses from 11,000 executives in 116 countries, asked participants to identify the top five risks most likely to affect their nations over the next two years from a list of 34 possible threats.

While economic issues such as inflation and downturns, along with societal challenges like polarization and inadequate public services, remained dominant, technology-driven risks stood out prominently in this year’s results.

Within G20 nations, concerns over AI were especially visible. “Adverse outcomes of AI technologies” emerged as the leading risk in Germany and the fourth most significant in the US. Australian executives similarly flagged “adverse outcomes of frontier technologies,” including quantum innovations, as a top threat.

Misinformation and disinformation ranked as the third-largest concern for executives in the US, UK and Canada. Meanwhile, in India, cyber insecurity—including threats to critical infrastructure—was identified as the number one risk.

Regionally, mis/disinformation ranked second in North America, third in Europe and fourth in East Asia. Cyber insecurity was the third-highest risk in Central Asia, while concerns around harmful AI outcomes placed fourth in South-east Asia.

AI’s influence is clearly woven through most of the technological risks highlighted. The technology is enabling more sophisticated disinformation efforts, including realistic deepfake audio and video. At the same time, AI is heightening cyber risks by empowering threat actors with advanced capabilities in social engineering, reconnaissance, vulnerability analysis and exploit development, according to the UK’s National Cyber Security Centre (NCSC).

The NCSC’s recent threat outlook cautions that AI will “almost certainly” make several stages of cyber intrusion “more effective and efficient” in the next few years.

The survey’s references to “adverse outcomes” of AI also include potential misuse of agentic or generative AI tools and the manipulation of AI models for disruptive, espionage-related or malicious purposes.

A study released in September found that 26% of US and UK organizations experienced a data poisoning attack in the past year, underscoring the growing risks.

“With the rise of AI, the proliferation of misinformation and disinformation is enabling bad actors to operate more broadly,” said Andrew George, president of Marsh Specialty. “As such, the challenges posed by the rapid adoption of AI and associated cyber threats now top boardroom agendas.”