Recent research shows that as businesses adopt AI more deeply into their workflows, new data risks are emerging. Companies are experimenting with AI tools across development, operations, and knowledge work, but many lack the visibility needed to manage these risks effectively. A new report by Cyberhaven Labs sheds light on how AI use is evolving and why organizations need to improve their data governance strategies.

The Growing Divide in AI Adoption

The report finds that AI adoption isn’t happening uniformly across organizations. Instead, it’s becoming more polarized. The top 1% of early adopters are using hundreds of Generative AI tools, while most companies remain cautious, using fewer than 15. This gap means some organizations push AI technology rapidly, while others lag behind, creating different levels of risk and security challenges.

This divide makes it harder for businesses to develop consistent policies around AI use. Companies that move fast may introduce more vulnerabilities, especially if security measures don’t keep up. Meanwhile, those more hesitant might miss out on AI benefits but also avoid certain risks. Understanding these differences is key for managing AI safely in the workplace.

Risks in Common AI Tools and Employee Data Behavior

The report highlights that many AI tools used today are risky from a security standpoint. Most employees enter sensitive data into these tools regularly, often without realizing the dangers. In the top AI applications, more than 80% are rated as medium or higher risk, yet employees still frequently use them to share confidential information.

Cyberhaven’s data shows that a significant portion of AI usage happens through personal accounts, which makes monitoring harder. Almost one-third of all data movements into AI tools involve sensitive information, including prompts and copy-paste actions. This behavior limits organizations’ visibility into what data is flowing where, making it difficult to control or protect sensitive information effectively.

Without proper oversight, organizations risk data leaks, compliance issues, and loss of trust. As AI tools become more embedded in daily work, understanding how data moves and ensuring controls are in place is more urgent than ever.

The Rise of Coding Assistants and AI Agents in the Workplace

Another key finding is the growth of coding assistants and AI agents as the “second wave” of workplace AI. Tools like GitHub Copilot and Claude Code are increasingly used by developers to write and review code more efficiently. In organizations leading in AI adoption, nearly 90% of developers use these tools, compared to about 50% in average companies.

This widespread use marks a shift in how AI is impacting technical work. These assistants help speed up development but also introduce new risks if sensitive code or data is shared with AI tools without proper safeguards. As AI becomes more integrated into daily tasks, companies need to ensure their security policies evolve accordingly.

Overall, the report emphasizes that AI’s rapid growth in workplaces requires a proactive approach to data governance. Without better visibility and controls, organizations could face significant security and compliance challenges as they embrace AI innovations.

Inspired by

• https://ai-techpark.com/cyberhaven-research-highlights-urgent-need-for-ai-data-governance/