Moving targets: Cybersecurity in a dynamic digital world

Cyber risk is becoming more complex, with artificial intelligence now the dominant concern. PwC’s 2027 Global Digital Trust Insights survey finds that AI-system attacks are the leading global security threat, cited by 50% of respondents. These threats cut across regions and demand immediate attention and investment. Frontier AI models are also raising the stakes by testing assumptions on which existing security programmes were built. The pace of change is itself a core cybersecurity challenge.

Cloud threats rank second globally, identified by 40% of leaders. As organisations accelerate digital transformation and move more infrastructure to the cloud, their potential exposure expands while attacks continue to evolve. This reinforces the need for robust cloud-security controls and governance to protect critical assets and data.

Third-party breaches complete the top three at 34% and have remained among the leading threats for four consecutive years. Many organisations still lack visibility and control across their extended ecosystems, leaving them exposed to breaches through trusted suppliers and partners. Regulations such as the Digital Operational Resilience Act (DORA) and the Network and Information Security Directive 2 (NIS2) are increasing the focus on third-party risk management.

AI-system attacks, cloud threats, and third-party breaches now define the leading cyber challenges for businesses. Ransomware, quantum-computing exposure, and distributed denial-of-service (DDoS) attacks also remain prominent. AI’s emergence as the leading concern reflects both its transformative potential and its growing security implications.

Managing these interconnected risks requires an integrated cybersecurity strategy that balances innovation with rigorous controls and visibility across operations. Yet only 17% of Irish respondents report that cyber risk is integrated into their organisation’s enterprise risk management process, compared with 41% globally. Closing that gap can support clearer ownership and better-informed investment decisions as organisations respond to fast-moving threats.

Organisations are reshaping their cybersecurity strategies as threats intensify, regulation expands, and operational resilience becomes a business priority. The shift is visible not only in larger budgets, but also in where organisations are directing investment.

Globally, 84% of organisations plan to increase their cyber budgets in 2027. Even amid economic pressure, the planned increases indicate that cyber resilience remains an investment priority for many organisations.

AI security is now the leading investment priority for the next 12 months, cited by 58% of organisations globally and 53% in Western Europe. Cloud security ranks second, at 46% globally and 44% in Western Europe, followed by data protection and trust, at 39% and 32%, respectively. This rebalancing reflects the growing prominence of AI-driven threats in organisational risk portfolios.

Investment, however, is only part of the response. Just 39% of organisations have fully formalised and integrated cyber continuity plans, with 23% of respondents globally not developing formal operational continuity plans related to cybersecurity. Third-party resilience is an even broader challenge, with 91% of organisations globally reporting material obstacles. These include enforcing service-level agreements, planning for resilience, exit, and portability, and understanding vendor concentration risk.

Higher spending will not close these gaps on its own. Organisations also need clearer third-party contracts, stronger vendor-management practices, and better visibility of supply chain risks. Pairing investment with these structural improvements will help turn cyber spending into stronger operational resilience.

AI-powered attacks are reshaping the threat landscape faster than many organisations can adapt. Despite increased investment and greater strategic awareness, gaps in preparedness leave businesses exposed to risks that are growing in speed, scale, and sophistication.

Frontier AI is at the centre of this shift. Autonomous botnets are the leading global concern, cited by 53% of respondents. Adversarial attacks and data poisoning follow at 52%, with adaptive malware and polymorphic evasion at 47%. These techniques can enable attacks to adapt and evade traditional controls, requiring organisations to rethink how they detect and contain them.

Organisations are responding by investing in AI-powered cyber defence. Threat detection and alerting are the leading planned uses, cited by 50% globally and 46% in Western Europe, followed by fraud detection and phishing detection and response. Cyber defence is moving from reactive response towards faster detection, prioritisation, and intervention.

Responsible AI governance and compliance lead cyber-AI investment at 42% globally. This indicates that accountability and regulatory alignment are prominent investment priorities alongside the development of more advanced defensive capabilities.

The readiness gap is particularly pronounced in Ireland. Only 34% of Irish organisations are investing in mitigating AI-enabled attacks, compared with 46% globally. Separately, 37% are investing in securing AI agents, compared with 45% globally. Preparation for quantum-computing threats is weaker still. Just 6% of Irish respondents are implementing quantum-resistant security measures, compared with 21% globally and 18% in Western Europe. A further 37% are exploring options without concrete implementation steps, while 20% have not begun to consider the risk.

Closing this gap requires coordinated action: strengthen defences against AI-powered attacks, improve incident-response capabilities, and begin the transition to quantum-resistant security. Without progress on all three fronts, Irish organisations may struggle to keep pace with an increasingly complex threat environment.

Autonomous AI agents are emerging as important tools in cyber defence, but Irish organisations are allowing them less operational autonomy than their global peers. Concerns about reliability, security, and regulation are shaping a preference for greater human oversight. That caution may protect against poorly governed deployment, but it could also slow threat response and limit operational efficiency.

Only 6% of Irish organisations allow AI agents to act fully autonomously in cyber defence, compared with 22% globally and 24% in Western Europe. Instead, 37% favour a “recommend, then confirm” model, compared with 28% globally and 26% in Western Europe. A further 11% do not use or plan to use autonomous agents, compared with 2% globally.

These choices reflect substantive concerns. Reliability and maturity are barriers for 69% of Irish respondents, compared with 55% globally and in Western Europe. Adversarial manipulation concerns 57%, while the same proportion cite regulatory constraints, well above the global rate of 36%.

Despite these concerns, global interest in autonomous agents is high. Some 96% of organisations plan to deploy autonomous agents in at least one security domain within the next 12 months. Threat-intelligence enrichment leads planned deployment at 49%, followed by phishing quarantine at 47% and malware removal at 46%. As adoption expands, the question is how to build the governance and trust needed to use these tools responsibly.

PwC’s recent AI Performance survey connects these foundations with the wider returns on AI. It found that 20% of companies capture three-quarters of AI’s economic gains. These AI leaders are 1.7 times as likely as other companies to have a Responsible AI framework and 1.5 times as likely to have a cross-functional AI governance board. Their employees are also twice as likely to trust AI outputs.

For Irish organisations, the priority is to make responsible AI governance part of adoption from the outset. That means establishing clear accountability alongside improving AI reliability and maturity, strengthening defences against adversarial manipulation, and translating regulatory requirements into clear operating guardrails. Together, these measures can help build trust and support faster threat response and more efficient cyber operations while retaining appropriate human control.