PwC Global Survey: Half of Executives Feel Unprepared for Attacks on AI Systems

PwC's 2027 Global Digital Trust Insights survey, covering 3,934 executives across 71 countries, reveals that AI has emerged as the primary cybersecurity challenge for enterprises. Among 1,819 surveyed AI and security leaders, half stated that attacks on AI systems are the hardest crisis to prepare for, outpacing cloud threats (40%), third-party breaches (34%), and ransomware (33%). Furthermore, when asked about AI-driven threats they feel least prepared for over the coming year, the highest responses were autonomous botnets (53%), adversarial attacks (52%), and data poisoning (52%). To counter evolving threats, 84% of respondents expect cybersecurity budgets to grow, with 58% prioritizing AI investments in governance, compliance, platform hardening, and supply chain security. However, readiness gaps persist: 49% of organizations have not yet prepared for quantum security, and only 39% have formal operational continuity plans for cyber disruptions.

Key takeaways

• Among 1,819 surveyed leaders, half consider attacks targeting AI systems the hardest crisis to prepare for.
• Autonomous botnets (53%), adversarial attacks (52%), and data poisoning (52%) represent the AI-driven threats executives feel least prepared for over the coming year.
• While 84% expect security budgets to rise and 58% prioritize AI investments, governance and compliance remain the top focus.
• Readiness gaps persist, with 49% yet to begin quantum security preparation and only 39% holding formal continuity plans for cyber disruptions.