The average global cost of a data breach has risen 12 percent year-on-year to a record $4.99 million, as businesses face growing financial losses from increasingly sophisticated cyberattacks, according to IBM’s 2026 Cost of a Data Breach Report.
The latest figure represents a significant increase in the financial burden associated with cyber incidents, with detection and escalation costs, as well as losses from disrupted business operations, accounting for 63 percent of the total cost of breaches recorded in the report.
IBM said the findings underscore the growing need for organisations to strengthen their cybersecurity systems as attackers increasingly use artificial intelligence to accelerate and scale attacks.
The report found that AI-driven attacks increased by 56 percent over the past year, with more than one in four organisations that experienced a malicious attack reporting that the incident was driven by AI.
Such attacks also carried a higher financial cost. Organisations affected by AI-driven attacks recorded an average breach cost of $6.04 million, about $1 million more than the $5.03 million average cost recorded for malicious attacks that did not involve AI.
IBM said cybercriminals are increasingly using AI to automate attacks, identify vulnerabilities and generate convincing malicious content, allowing them to operate at a speed and scale that would be difficult to achieve manually.
“Attackers are abandoning human speed for machine speed,” IBM said, noting that generative AI has lowered the time, cost and expertise required to launch sophisticated attacks.
Deepfake impersonation accounted for 45 percent of AI-driven attacks identified in the report, making it the most common form of AI-enabled attack. AI-generated malware accounted for 19 percent, while AI-generated phishing and other communications represented 17 percent.
AI systems become new attack surfaces
Beyond using artificial intelligence to facilitate cyberattacks, criminals are increasingly targeting the AI systems themselves as businesses deploy the technology across their operations.
IBM found that 21 percent of organisations experienced a security incident involving an AI model or application, up from 13 percent a year earlier.
The most expensive AI-related security incidents were model inversion and prompt injection attacks, which resulted in average losses of $6.07 million and $5.89 million respectively.
Model inversion attacks attempt to extract sensitive information from AI models, while prompt injection attacks seek to manipulate AI systems through malicious instructions.
IBM warned that the risks are becoming more significant as AI models become increasingly integrated into corporate applications, sensitive data and business workflows.
The report also found that 92 percent of organisations that experienced an AI-related breach lacked proper AI access controls, despite identity and access management being among the measures capable of significantly reducing breach costs.
Only 40 percent of organisations reported having access controls in place for AI models and data, leaving businesses exposed to additional attack paths.
Ransomware remains a major threat
Ransomware incidents also increased during the period covered by the report, accounting for 39 percent of breaches among affected organisations, compared with 34 percent in the previous year.
While data encryption remained a major ransomware tactic, IBM said attackers were increasingly turning to other forms of extortion to pressure victims into paying.
Threats to damage an organisation’s brand reputation emerged as the most common ransomware tactic, reported by 41 percent of organisations that experienced ransomware incidents, while encryption of data and systems accounted for 23 percent.
The increasing use of AI is also contributing to the evolution of ransomware, with criminals using the technology to generate malware and expand their methods of extortion.
AI offers cybersecurity savings
Despite the growing threat posed by AI-enabled attacks, IBM’s report also found that organisations using AI and automation extensively in their security operations were able to reduce the impact of breaches.
Security teams that extensively deployed AI and automation shortened breach response times by 65 days and reduced average breach costs by $1.93 million compared with organisations that did not extensively use the technologies.
However, adoption across the security lifecycle remains uneven.
Only 36 percent of breached organisations said they extensively used AI and automation across prevention, detection, investigation and response. IBM said the technologies are more commonly deployed for detection and investigation, while their use in prevention remains comparatively limited.
The report also found that among organisations deploying AI agents for cybersecurity, more than half were using them for threat hunting, response and containment, while only 18 percent applied them to vulnerability scanning and management.
IBM said greater use of AI in preventive security could help organisations identify vulnerabilities earlier and reduce their exposure to emerging threats.
Africa’s cybercrime losses rise
The global increase in cyber risks comes as African organisations also face a rapidly evolving threat landscape.
The 2026 African Cyberthreat Assessment Report by INTERPOL found that artificial intelligence was enabling 55 percent of reported cybercrimes across the continent.
Financial losses linked to cybercrime more than doubled, rising from $192 million in 2024 to $484 million in 2025.
INTERPOL attributed the development to the increasing industrialisation of cybercrime, with criminal networks moving beyond isolated attacks to more organised, borderless operations supported by AI tools.
