Artificial intelligence (AI) adoption is greatly outpacing AI security and governance.
IBM’s new Cost of a Data Breach Report suggests that AI is already being seen as an easy, high-value target by threat actors.
While the overall number of organisations experiencing an AI-related breach is a small representation of the researched population, this is the first time security, governance and access controls for AI have been studied in this report.
Although 13% of organisations reported breaches of AI models or applications, 8% of organisations reported not knowing if they had been compromised in this way. And, of those compromised, 97% report not having AI access controls in place.
As a result, 60% of the AI-related security incidents led to compromised data and 31% led to operational disruption.
This year’s results show that organisations are bypassing security and governance for AI in favour of do-it-now AI adoption. Ungoverned systems are more likely to be breached – and more costly when they are.
“The data shows that a gap between AI adoption and oversight already exists, and threat actors are starting to exploit it,” says Suja Viswesan, vice-president: security and runtime products at IBM. “The report revealed a lack of basic access controls for AI systems, leaving highly sensitive data exposed, and models vulnerable to manipulation.
“As AI becomes more deeply embedded across business operations, AI security must be treated as foundational. The cost of inaction isn’t just financial, it’s the loss of trust, transparency and control.”
However, the report did reveal that organisations using AI and automation extensively throughout their security operations saved an average $1,9-million in breach costs and reduced the breach lifecycle by an average of 80 days.
The 2025 report, conducted by Ponemon Institute, sponsored and analysed by IBM, is based on data breaches experienced by 600 organisations globally from March 2024 through February 2025.
Key findings from the report around AI security and breaches, the financial cost of a breach, and operational disruption are as follows:
Breaches and the AI era
- AI governance policies – 63% of breached organisations either don’t have an AI governance policy or are still developing a policy. Of the organisations that have AI governance policies in place, only 34% perform regular audits for unsanctioned AI.
- The cost of shadow AI – One in five organisations reported a breach due to shadow AI, and only 37% have policies to manage AI or detect shadow AI. Organisations that used high levels of shadow AI observed an average of $670 000 in higher breach costs than those with a low level or no shadow AI. Security incidents involving shadow AI led to more personally identifiable information (65%) and intellectual property (40%) being compromised compared to the global average (53% and 33% respectively).
- Smarter Attacks with AI – 16% of breaches studied involved attackers using AI tools, most often for phishing or deepfake impersonation attacks.
The financial cost of a breach
- Data breach costs – The global average cost of a data breach fell to $4,44-million, the first decline in five years, while the average US cost of a breach reached a record $10,22-million.
- Global breach lifecycles hit record low – The global average breach lifecycle (the mean time to identify and contain a breach, including restore services) dropped to 241 days, a 17-day reduction from the year prior, as more studied organisations detected the breach internally. Those organisations who detected the breach internally also observed a $900 000 savings on breach costs compared to those disclosed by an attacker.
- Healthcare breaches remain the costliest – Averaging $7,42-million, healthcare breaches remained the most expensive across all studied industries, even as this sector saw a $2,35-million reduction in costs compared to 2024. Breaches across this sector take the longest to identify and contain at 279 days, that’s more than five weeks longer than the global average of 241 days.
- Ransom payment fatigue – Last year, organisations pushed back against ransom demands, with more opting not to pay (63%) compared to the year prior (59%). As more organisations refuse to pay ransoms, the average cost of an extortion or ransomware incident remains high, particularly when disclosed by an attacker ($5,08-million).
- Security investments stall amid rising AI risks – There was a significant reduction in the number of organisations that said they plan to invest in security following a breach, 49% in 2025 compared to 63% in 2024. Less than half of those that plan to invest in security post-breach will focus on AI-driven security solutions or services.
The long tail of a breach: operational disruption
According to the 2025 IBM report, nearly all organisations studied suffered operational disruption following a data breach. This level of disruption is taking a toll on recovery timelines.
Among organisations that reported recovery, most took more than 100 days on average to do so.
However, the consequences of a breach continue to extend beyond containment. While down compared to the year prior, nearly half of all organisations reported that they planned to raise the price of goods or services because of the breach, and nearly one-third reported price increases of 15% or more.
Changes over the years
The Cost of a Data Breach Report has investigated nearly 6 500 data breaches over the past 20 years. Since the inaugural report in 2005, the nature of breaches has evolved dramatically. Back then, risk was largely physical. Today, the threat landscape is overwhelmingly digital and increasingly targeted, with breaches now driven by a spectrum of malicious activity.
With the pace of enterprise AI adoption proliferating, for the first time, the Cost of a Data Breach research studied the state of security and governance for AI, the type of data targeted in security incidents involving AI, breach costs associated with AI-driven attacks, and the prevalence and risk profile of shadow AI (unregulated, unauthorised use of AI).
Historical findings from past reports include the following:
- 2005: Nearly half (45%) of all data breaches were caused by lost or stolen computing devices, such as a laptop or thumb drive, and only 10% of breaches were due to hacked electronic systems.
- 2015: Breaches due to cloud misconfiguration weren’t even a categorised threat; today they are a leading target.
- 2020: Ransomware began to surge, and by 2021 it accounted for an average of $4,62-million in breach costs, and this year that number reached an average of $5,08-million (when the incident was disclosed by an attacker).
- 2025: AI, which was included for the first time in the research this year, is quickly emerging as a high-value target.