By 2027, more than 40% of AI-related data breaches will happen due to generative AI (GenAI) misuse across borders, predicts Gartner. The fast growth of GenAI tools has gone beyond the ability to create rules and security measures for data. This has raised worries about needing to keep data local since GenAI needs a lot of computer power that isn’t always in the same place as the data.

Global AI Rules Are Not the Same, Leading to Problems

The lack of clear global rules and best practices for AI and data safety makes things harder. Companies have to come up with different plans for each region, which can stop them from growing and taking full advantage of AI products and services.

Companies Need to Take Action Before AI Rules Are Required Everywhere Gartner believes that by 2027, every country will need to have rules for AI.

To reduce the risks of data problems, especially from cross-border misuse of GenAI, and to follow the rules, Gartner suggests several steps for companies:

1. Improve Data Management: Companies should follow international rules and keep an eye on any unintended data transfers between countries. They need to include guidelines for data that AI processes in their data management plans.

2. Create Governance Teams: Companies should set up teams to oversee AI use and ensure clear communication about how they use AI and handle data. These teams will look after technical issues and manage risks and compliance.

3. Boost Data Security: Companies should use advanced technologies, encryption, and anonymization to keep sensitive data safe. They should verify secure computing environments in specific areas and use advanced methods like Differential Privacy to protect data when it needs to be shared.

4. Invest in Trust and Security Tools: Companies should plan budgets for tools that help manage trust, risk, and security related to AI. This includes governance for AI, data security management, and controlling data output. Gartner predicts that by 2026, companies using these controls will handle at least 50% less incorrect or illegitimate information, helping them make better decisions.

Source: https://www.gartner.com/en/newsroom/press-releases/2025-02-17-gartner-predicts-forty-percent-of-ai-data-breaches-will-arise-from-cross-border-genai-misuse-by-2027