AI Security Risks Rise as Multi-Domain Attacks Increase in 2026

70% of companies have experienced a security breach due to AI-related vulnerabilities, highlighting the growing need for robust AI security measures. According to recent studies, the furious pace of AI adoption is putting security progress at risk, with businesses moving fast to self-host LLM infrastructure without adequately addressing security concerns.

Understanding AI Security Risks

The recent funding of $3 million by Herd Security to expand its AI capabilities is a step in the right direction, but more needs to be done to address the looming security threats. 1 million exposed AI services have been found to have significant security vulnerabilities, posing a substantial risk to organizations. Furthermore, 60% of AI tools and workflows are connected to Google or Microsoft, leaving behind persistent OAuth tokens that can be exploited by attackers.

OAuth Token Vulnerabilities

• 90% of organizations do not have adequate measures in place to monitor and clean up OAuth tokens, leaving them vulnerable to attacks.
• 50% of AI-powered applications use OAuth tokens that have no expiration date, further increasing the risk of security breaches.

"The security risks associated with AI are real and pressing. It's essential for organizations to take proactive measures to address these risks and ensure the secure adoption of AI technologies." - Peter Sarlin, AI Security Expert

What the Sceptics Say

Some sceptics argue that the emphasis on AI security risks is overstated and that the benefits of AI adoption far outweigh the potential drawbacks. However, this perspective neglects the potential financial losses that can result from AI-related security breaches, which can be substantial.

What This Means for the Industry

Companies like Boost Security and Herd Security are at the forefront of addressing AI security risks. In the next 6-12 months, we can expect to see a significant increase in investment in AI security solutions, with a focus on multi-domain attack prevention and OAuth token management. Google and Microsoft will also need to take steps to address the security vulnerabilities associated with their platforms.

Key Takeaways

1. Engineers: Prioritize the implementation of robust security measures when developing AI-powered applications, including the use of secure OAuth tokens and regular security audits.
2. Investors: Invest in companies that are developing innovative AI security solutions, such as Herd Security and Boost Security.
3. Business Leaders: Ensure that your organization has a comprehensive AI security strategy in place, including regular security assessments and employee training.
4. Consumers: Be aware of the potential security risks associated with AI-powered applications and take steps to protect your personal data, such as using strong passwords and enabling two-factor authentication.

Engineers should immediately review their AI application's security measures, investors should consider investing in AI security startups, and business leaders should develop a comprehensive AI security strategy. Consumers should also take proactive steps to protect their personal data.

Sources

• SiliconANGLE: Security training startup Herd Security raises $3M to expand AI capabilities
• The Hacker News: The Back Door Attackers Know About — and Most Security Teams Still Haven’t Closed
• Krebs on Security: How AI Assistants are Moving the Security Goalposts