AI Risks
We've curated 17 cybersecurity statistics about AI Risks to help you understand how emerging threats like deepfakes and automated attacks are reshaping the landscape of cybersecurity in 2025.
Related Topics
Showing 1-17 of 17 results
48% of UK consumers express concern about the risk of fraud or identity theft related to AI in banking.
57% of organizations reported lacking the ability to block risky AI actions in real time.
66% of organizations reported catching AI tools over-accessing sensitive information.
60-70% of AI-generated code lacks deployment environment awareness, generating code that runs locally but fails in production.
40-50% of AI-generated code inflates coverage metrics with meaningless tests rather than validating logic.
80-90% of AI-generated code rigidly follows conventional rules, missing opportunities for more innovative, improved solutions.
80-90% of AI-generated code creates hyper-specific, single-use solutions instead of generalizable, reusable components.
80-90% of AI-generated code generates functional code for immediate prompts but never refactors or architecturally improves existing code.
70-80% of AI-generated code violates code reuse principles, causing identical bugs to recur throughout codebases, requiring redundant fixes.
40-50% of AI-generated code reimplements from scratch instead of using established libraries, SDKs, or proven solutions.
20-30% of AI-generated code over-engineers for improbable edge cases, causing performance degradation and resource waste.
90-100% of AI-generated code contains excessive inline commenting, which dramatically increases computational burden and makes code harder to check.
40-50% of AI-generated code defaults to tightly-coupled monolithic architectures, reversing decade-long progress toward microservices.
The percentage of companies globally that felt very prepared to manage AI risks has remained relatively flat over the past three years, with 9% in 2023, 8% in 2024, and 12% in 2025.
94% of all AI services are at risk for at least one of the top Large Language Model (LLM) risk vectors, including prompt injection/jailbreak, malware generation, toxicity, and bias.
30% of respondents report the emergence of a new attack surface due to the use of AI by their business users.
Approximately 1 in 4 organizations said they’re concerned about how AI use in the enterprise will make them more attackable (AI and generative AI concerns)