AI cyberattacks surge 56% as average breach cost hits $6.04m-IBM

AI-driven cyberattacks increased by 56 per cent in the past year, with more than one in four organisations experiencing a malicious attack involving artificial intelligence, as the average cost of an AI-driven sbreach rose to $6.04 million, according to IBM.

The findings, contained in IBM’s 2026 Cost of a Data Breach Report, show that AI is rapidly changing the economics of cybercrime by allowing attackers to automate operations, identify vulnerabilities faster and launch sophisticated attacks at a much larger scale.

IBM said AI-driven attacks now cost businesses about $1 million more per breach than malicious attacks that do not involve AI, with the average cost rising to $6.04 million from $5.03 million.

The sharp increase has significant implications for businesses as artificial intelligence becomes more deeply embedded in corporate systems, financial platforms and operational processes.

More than 25 per cent of organisations surveyed experienced a malicious AI-driven attack during the period, representing a 56 per cent increase from the previous year.

Deepfake impersonation was the most common form of AI-driven attack, accounting for 45 per cent of incidents identified by IBM.

AI-generated malware represented another 19 per cent, while AI-generated phishing and other communications accounted for 17 per cent.

IBM said generative AI has reduced the cost, time and expertise required to execute social-engineering attacks, enabling criminals to produce convincing fraudulent messages and impersonate individuals at greater scale.

“Attackers are abandoning human speed for machine speed,” IBM said.

The threat is also moving in the opposite direction, with cybercriminals increasingly targeting the artificial intelligence systems businesses are deploying.

IBM found that 21 per cent of organisations experienced a security incident involving an AI model or application, up from 13 per cent a year earlier.

The most expensive attacks against AI systems involved model inversion, with an average cost of $6.07 million per incident, and prompt injection, which averaged $5.89 million.

Model inversion involves attempts to extract sensitive information from an AI model, while prompt injection seeks to manipulate an AI system through malicious instructions.

The findings highlight an emerging double risk for businesses: AI is being used both as a tool for attacking organisations and as a new target for cybercriminals.

The trend is also becoming increasingly costly across Africa.

According to INTERPOL’s African Cyberthreat Assessment Report 2026, artificial intelligence was linked to 55 per cent of reported cybercrimes across the continent, while financial losses from cybercrime more than doubled from $192 million in 2024 to $484 million in 2025.

The report, based on survey data from 36 African member countries, said cybercrime was increasingly shifting from isolated attacks to large-scale, borderless criminal networks powered by AI tools.

For businesses, the economic implications extend beyond the immediate cost of recovering compromised systems. Data theft, operational disruption, regulatory penalties, reputational damage and lost customer confidence can significantly increase the total cost of an attack.

The United States has also recorded substantial financial losses from AI-enabled cybercrime. The FBI reported more than $893 million in AI-related cybercrime losses in 2025, reflecting the growing use of synthetic voices, fake identities, manipulated videos and other AI-generated content in fraud.

The accelerating use of AI by cybercriminals therefore creates a growing cost for businesses at precisely the time many organisations are increasing their reliance on artificial intelligence.

For Nigerian and African businesses, the figures point to the need for stronger cybersecurity investment, tighter controls around AI systems and greater employee awareness as companies adopt AI across payments, finance, customer service and other business operations.

The economic risk is increasingly clear: as AI reduces the cost and speed of launching cyberattacks, a single successful breach can impose a multimillion-dollar financial burden on a business.