Essential data on market growth, security vulnerabilities, governance gaps, and regulatory trends shaping how organizations build and deploy trustworthy AI systems
AI trust and safety has shifted from an afterthought to a business imperative. As AI systems handle increasingly sensitive tasks, from processing personal photos to managing financial decisions, the stakes around data privacy, security, and ethical use continue to rise. The statistics below reveal how rapidly this landscape evolved through 2025, where the biggest gaps remain, and what organizations are doing to build more trustworthy AI.
Whether you are evaluating AI tools for personal use, building AI-powered products, or setting governance policies at your organization, these numbers provide essential context for understanding the current state of AI privacy and data concerns.
Key Takeaways
- Market growth is accelerating: The global AI Trust and Safety market grew from $8.2 billion in 2024 and is projected to reach $39.2 billion by 2033, representing nearly fivefold growth over the forecast period
- Shadow AI creates costly breaches: Organizations with high levels of shadow AI faced breach costs averaging $670,000 more than those with little or no shadow AI
- Governance gaps persist: 63% of organizations still had no AI governance policy in place, while only 23% had implemented AI runtime controls
- Trust remains fragile: 70% of Americans have little to no trust in companies to make responsible AI decisions
- AI incidents are increasing: Documented AI-related incidents reached a record 233 in 2024, up 56.4% from the previous year
- AI-powered security delivers ROI: Organizations using AI extensively in security cut average breach costs by up to $1.9 million
- Research focus is shifting: Responsible AI papers at leading conferences increased 28.8% year-over-year
The Global Landscape of AI Usage and Growth Statistics
The AI Trust and Safety market grew substantially through 2025 as organizations recognized that deploying AI responsibly requires dedicated investment in governance, risk management, and security infrastructure.
1. The global AI Trust and Safety market reached $8.2 billion in 2024 with a path to $39.2 billion by 2033
The AI Trust and Safety sector experienced rapid expansion. The market was valued at $8.2 billion in 2024 and is projected to reach $39.2 billion by 2033, representing nearly fivefold growth over the forecast period. This growth reflected increasing enterprise demand for tools that ensure AI systems operate safely, fairly, and transparently.
2. AI Trust, Risk, and Security Management reached $2.95 billion in 2025
A closely related market segment, AI Trust, Risk, and Security Management (AI TRiSM), reached $2.95 billion in 2025 and is projected to hit $21.06 billion by 2035 at a 21.72% CAGR. This segment focuses specifically on managing the risks associated with AI model deployment, including security vulnerabilities, compliance gaps, and governance frameworks.
3. North America led with $3.2 billion in AI Trust and Safety revenue
Regional adoption varied significantly. North America accounted for approximately $3.2 billion in AI Trust and Safety market revenue in 2024, driven by enterprise AI adoption and regulatory pressure from federal agencies.
4. Asia Pacific AI Trust and Safety market reached $2.1 billion with the fastest growth rate
The Asia Pacific region captured $2.1 billion of the market in 2024, with the highest projected growth rate at 22.1% CAGR through 2033. Rapid digital transformation across manufacturing, finance, and consumer technology fueled this expansion.
Understanding AI’s Potential Dangers: Key Statistics on AI Risks
As AI adoption accelerated through 2025, so did the risks. From data breaches to algorithmic bias, understanding where AI systems fail is essential for anyone building or using these technologies.
5. AI-related incidents reached a record 233 in 2024, up 56.4% from the prior year
The number of documented AI-related incidents rose to 233 in 2024, marking a record high and a 56.4% increase over 2023. These incidents spanned categories including privacy violations, algorithmic bias, security exploits, and harmful outputs from generative AI systems.
6. 20% of surveyed organizations reported a breach involving shadow AI
Unsanctioned AI tools, often called shadow AI, represented a significant security risk. In IBM’s study of 600 organizations, 20% reported experiencing a breach involving unsanctioned or shadow AI tools. Employees adopting consumer AI tools without IT oversight created blind spots that traditional security measures failed to catch.
7. Organizations with high shadow AI levels faced $670,000 higher breach costs
When shadow AI was involved in a breach, the financial impact was severe. Organizations with high levels of shadow AI faced breach costs averaging $670,000 more than organizations with little or no shadow AI. This premium reflected the difficulty of detecting, containing, and remediating breaches when the AI tools involved were never properly inventoried or secured.
8. 97% of organizations with AI security incidents lacked proper access controls
Among organizations that experienced AI-related security incidents, 97% lacked proper AI access controls. Without clear policies governing who can deploy AI tools and what data those tools can access, organizations exposed themselves to preventable risks.
9. 13% of surveyed organizations reported a breach involving their AI models or applications
IBM found that 13% of surveyed organizations reported a breach involving their AI models or applications, often through compromised applications, APIs, plug-ins, or other supply-chain components.
AI Ethics: Statistics on Trustworthy AI Principles
Ethical AI is not just an abstract principle. Organizations increasingly measured and reported on fairness, transparency, and accountability as core business metrics through 2025.
10. 64% of organizations identify inaccuracy as a top Responsible AI risk
When asked about their primary Responsible AI concerns, 64% of organizations cited inaccuracy as a key risk. Additionally, 63% expressed concerns about regulatory compliance, and 60% listed cybersecurity threats. These three risks dominated enterprise AI governance discussions.
11. 78% of consumers believe organizations must use AI ethically
Consumer expectations were clear: 78% of consumers believe organizations have a responsibility to use AI ethically. Companies that failed to meet this expectation risked losing consumer trust and facing regulatory scrutiny.
12. Foundation model transparency scores improved from 37% to 58%
Transparency among major AI model developers improved. The average transparency score increased from 37% in October 2023 to 58% in May 2024. This metric tracked how openly developers disclosed information about training data, model capabilities, safety testing, and limitations.
13. Responsible AI research papers increased 28.8% at leading conferences
Academic focus on AI safety and ethics intensified. The number of Responsible AI papers accepted at leading AI conferences grew 28.8%, from 992 in 2023 to 1,278 in 2024. This research pipeline fed into practical tools and frameworks for safer AI deployment.
AI Security Vulnerabilities: Statistics on Generative AI and Agentic Systems
Generative AI and autonomous AI agents introduced new categories of security risk that traditional cybersecurity approaches struggled to address.
14. Among employees using generative AI for work, 69% also use personal tools through personal devices
Employee behavior created significant exposure. Among respondents who used generative AI for work, 69% said they also use their own tools through personal devices or accounts, while 57% use company-provided tools. This overlap created governance and security gaps when personal AI accounts were used for workplace tasks.
15. 47% of surveyed U.S. consumers experienced a digital security failure in the past year
Digital security failures remained common. Deloitte’s 2025 Connected Consumer Survey found that 47% of U.S. consumers experienced at least one incident in the past year, such as a hacked device, account breach, or stolen identity. This is a consumer cybersecurity statistic rather than an organization-level measure of AI incidents.
Building Trust: Statistics on User Concerns and AI System Reliance
Trust determined whether users adopted AI tools and how much value those tools could deliver. These statistics revealed significant gaps between what users expected and what they experienced.
16. 68% of global consumers are concerned about online privacy
Privacy remained a top-of-mind issue for most people. 68% of global consumers expressed concern about their privacy online, creating a challenging environment for AI tools that require access to personal data to function effectively.
17. 57% of consumers see AI as a significant threat to their privacy
Beyond general privacy concerns, 57% of consumers globally specifically agreed that AI posed a significant threat to their privacy. This perception shaped how people evaluated AI-powered products and services.
18. 61% of global respondents are wary about trusting AI systems
Trust in AI remained fragile. 61% of global respondents expressed wariness about trusting AI systems in general. Building this trust required transparency about how AI systems work, what data they access, and how that data is protected.
19. 70% of Americans have little to no trust in companies’ AI decisions
In the United States, skepticism ran particularly deep. 70% of Americans said they had little to no trust in companies to make responsible decisions about how they use AI in their products. This trust deficit created both challenges and opportunities for companies that could demonstrate genuine commitment to AI privacy.
Mitigating AI’s Dangers: Practical Approaches and Statistical Outcomes
Organizations that invested in AI governance and security saw measurable returns. These statistics showed the financial and operational benefits of proactive risk management.
20. Organizations using AI in security cut breach costs by up to $1.9 million
AI is not only a source of risk but also a powerful defense tool. Organizations using AI extensively in security operations reduced average breach costs by up to $1.9 million and shortened breach lifecycles by approximately 80 days.
21. Security AI and automation saved organizations $2.22 million on average
A separate analysis found that organizations deploying security AI and automation extensively achieved $2.22 million in average cost savings compared to those that did not. This ROI made a compelling case for investing in AI-powered security tools.
22. 63% of organizations have no AI governance policy in place
Despite the clear benefits of governance, 63% of organizations still lacked any AI governance policy. This gap left them exposed to shadow AI risks, compliance failures, and security incidents that proper policies could prevent.
23. Only 23% of organizations have implemented AI runtime controls
Even fewer organizations moved beyond policy to implementation. Just 23% had deployed AI runtime controls that actively monitor and govern how AI tools are used across the organization. This represented a significant opportunity for security and IT teams to close a known gap.
Implementation Priorities for AI Trust and Safety
Based on these statistics, organizations should consider the following priorities when building or evaluating AI systems:
Governance foundations:
- Establish clear AI governance policies before deployment
- Create inventory systems for all AI tools in use
- Define access controls and data handling requirements
Security measures:
- Implement runtime controls to monitor AI tool usage
- Address shadow AI through employee education and approved alternatives
- Deploy AI-powered security tools to detect and respond to threats
Trust-building practices:
- Prioritize transparency in how AI systems access and use data
- Consider on-device processing where feasible to minimize data exposure
- Communicate privacy practices clearly to users
Compliance readiness:
- Track regulatory developments in your operating jurisdictions
- Document AI system capabilities, limitations, and safety testing
- Build flexibility into governance frameworks to adapt to new requirements
Frequently Asked Questions
What are the most significant risks associated with artificial intelligence today?
The most pressing AI risks include security vulnerabilities from shadow AI. In IBM’s study of 600 organizations, 20% reported breaches involving unsanctioned AI tools, with organizations that had high levels of shadow AI facing breach costs averaging $670,000 more than those with little or no shadow AI. Beyond security, organizations cite inaccuracy (64%), regulatory compliance (63%), and cybersecurity (60%) as their top Responsible AI concerns. Privacy threats remain significant, with 57% of consumers viewing AI as a threat to their personal data.
How do companies ensure the ethical use and safety of their AI systems?
Effective AI governance combines policy, technology, and culture. This includes establishing clear governance frameworks, implementing access controls and runtime controls to monitor AI usage, and improving transparency around training data, system capabilities, safety testing, and limitations. However, 63% of organizations still had no AI governance policy, and only 23% had implemented AI runtime controls.
What statistics show that investment in AI trust and safety is growing?
The global AI Trust and Safety market was valued at $8.2 billion in 2024 and is projected to reach $39.2 billion by 2033. The related AI Trust, Risk, and Security Management market reached $2.95 billion in 2025 and is projected to grow to $21.06 billion by 2035. Responsible AI research also expanded, with papers accepted at leading conferences increasing 28.8% from 992 in 2023 to 1,278 in 2024.
How does privacy-focused design contribute to overall AI trust?
Privacy-focused design addresses the concerns of the 68% of global consumers who are worried about online privacy and the 57% who see AI as a significant privacy threat. Practices such as limiting data access, using clear access controls, improving transparency, and processing data on devices where feasible can help organizations address the trust gap that leaves 70% of Americans skeptical of companies’ AI decisions.
What can organizations do to reduce AI-related security risks?
Organizations can reduce AI-related risks by creating an inventory of approved AI tools, establishing governance policies, defining access controls, monitoring AI usage with runtime controls, and addressing employee use of personal AI accounts. These measures are especially important because 97% of organizations that experienced AI security incidents lacked proper access controls, while 69% of employees using generative AI for work also used personal tools through personal devices or accounts.