PWC’S Global Digital Trust Insights 2027 survey, which polled nearly 4,000 leaders across more than 70 countries, highlights a wide gap between the pace of AI and quantum threats and organisations’ preparations. The report notes a surge in AI-assisted attacks and identifies attacks on AI systems themselves as the top cyber threat leaders feel least prepared to defend against.
Specific concerns include compromise by autonomous botnets (53%), adversarial attacks (52%), and data poisoning (52%), with prompt injections—often described as prompt engineering—also recognised as a difficult problem, already noted by OpenAI as unlikely to be fully solved.
On the defence side, AI is seen as crucial for threat detection and alerting, yet adoption of fully autonomous AI for cyber defence remains limited: only 22% of leaders would utilise autonomous AI agents for cyber defence without human approval. The reluctance is driven by concerns over reliability and maturity (55%) and a reported skills gap in AI oversight and governance (44%).
The survey also emphasises that emerging quantum threats are driving a shift towards quantum-resistant cryptography, but only 21% of respondents are implementing quantum-resistant security measures. PwC warns of “harvest now, decrypt later” risk as quantum computers edge closer to breaking classical encryption with Shor’s algorithm, underscoring the need to replace encryption with quantum-resistant schemes.
The report signals a clear imbalance: while technology to strengthen cybersecurity exists, global adoption lags. Budget expectations reflect this, with 84% of security and finance leaders expecting increases and 58% naming AI as the top cyber budgeting priority.