CCI staff share recent surveys, reports and analysis on risk, compliance, governance, infosec and leadership issues. Share details of your survey with us: editor@corporatecomplianceinsights.com.
Using generative AI without permission is now the top employee security mistake
US and UK cybersecurity leaders are about as likely to blame employees’ unsanctioned AI use for their company’s worst security incident of the past year as they are phishing, according to a survey by information security and compliance platform IO.
The survey of 3,000 cybersecurity managers and executives found that, among those who experienced a security incident in the past 12 months, about the same share blamed shadow AI as blamed phishing or social engineering for their most serious incident or near miss: roughly a quarter each. Shadow AI includes giving an AI tool unapproved access to a business system and leaking confidential information through an AI tool.
Phishing, long the focus of security awareness training inside organizations, remains the most common type of incident respondents experienced (32%). But when it comes to the most serious incidents, shadow AI now rivals it. Using generative AI without permission was the most common employee security mistake, cited by 35% of respondents, and 22% said staff have built AI into important workflows without the security team knowing it.
Other key findings:
- Only 29% say their board collectively owns cyber risk, with a wide gap between the UK (34%) and US (21%).
- The share saying they are “very confident” in their ability to respond to a major incident fell from 61% to 41%.
- More than a quarter (27%) say AI security accountability is shifting from the security team to individual staff.
Almost 60% of legal leaders say AI adoption outpaces governance
More than half of legal leaders believe their organizations are adopting AI faster than they can govern the technology, according to a survey by Diligent Institute and law firm Wilson Sonsini.
The survey of about 200 leaders at businesses, nonprofits, government organizations and others found that 59% said governance is struggling to keep pace with AI adoption. More than two in five reported that their organizations lack a basic AI governance framework, with 14% reporting no formal approach and 27% still developing policies and guidelines. However, 38% have a basic framework in place and 21% said governance was established and regularly reviewed.
The lack of governance for many organizations comes as they worry about AI threats. Almost half (49%) said AI-related risks were among leading risks, an increase of 10 percentage points, from an earlier survey in 2026. But the changing regulatory environment was cited as the top risk factor, cited by 55%, up seven points. Geopolitical conflict fell sharply as a concern, down from 52% to 26%.
Other key findings include:
- Legal leaders rated their boards‘ readiness to tackle AI risks at 2.9 out of 5.
- Only 23% regularly brief boards or senior leaders on AI-related risks.
- Legal teams are actively involved in AI due diligence, with 67% saying they review vendors and assess privacy, data, confidentiality and intellectual property issues.
Data breach settlements cost organizations $3.2B since 2010
The cost of data breach settlements in federal class actions totaled $3.2 billion from 2010 to 2025, a period that saw a precipitous rise in these types of settlements, according to a study by Cornerstone Research.
The study looked at 426 federal class-action settlements over the period, with 357 having disclosed details, and found settlements and costs have dramatically risen and have weighed heaviest on financial services, technology and healthcare industries.
The lone data breach settlement from 2010 with a known amount was $13 million. From 2010-15, no more than six data breach settlements occurred each year, and none exceeded $50 million. Since 2016, the number of settlements has climbed, reaching 36 in 2022, 51 the next year, 91 in 2024 and 139 in 2025. Settlement dollars have risen too, peaking in 2020 at $889 million, driven by a $629 million settlement in a lawsuit against Equifax. Although high-value settlements of $50 million or more have slowed, with none in 2025, settlements that year still totaled $467 million.
Of the 347 settlements with monetary compensation, 76% were worth $5 million or less, 18% were between $5 million and $25 million, 3% were between $25 million and $50 million and another 3% were worth $50 million or more, the study found.
From 2010 to 2025, financial services paid the most for these settlements, at almost $1.2 billion. The tech industry paid $726 million, and healthcare paid $585 million, making it the third most burdened industry, though healthcare had the most settlements of any industry, with 118.









