IBM will utilize OpenAI’s cybersecurity models to help companies detect and assess software vulnerabilities more quickly. The collaboration is part of the OpenAI Daybreak Cyber Partner Program and builds on Project Lightwell, the security initiative for which IBM and Red Hat previously announced a $5 billion investment.
The new service focuses on application security and is designed to go beyond traditional code scanners. According to IBM, the solution analyzes application code using advanced AI models to identify potential vulnerabilities and attack vectors more quickly. It automatically prioritizes high-risk software components, enabling security teams to focus on the most urgent issues.
AI Within the Customer’s Own Environment
A key part of the approach is that the analyses take place within the customer’s IT environment. Through IBM Consulting Advantage, IBM Consulting’s AI platform, the models are granted read-only access to code repositories. This allows the AI to assess the software without modifying the source code.
IBM positions the service as a managed solution for enterprises. Organizations can start by assessing specific business-critical applications and later transition to continuous monitoring. Risks are reevaluated as soon as software changes or new threats emerge.
The collaboration builds on Project Lightwell, an initiative IBM unveiled last month. That project combines AI tools with a global team of engineers to inspect, repair, and manage open-source software in the supply chain. Within Lightwell, OpenAI’s cyber models are used alongside other advanced AI models for code review and vulnerability remediation.
According to Mark Hughes, head of cybersecurity services at IBM Consulting, attackers are now using AI to find and exploit vulnerabilities at scale and speed. That is why defenders need comparable technology, combined with the control and security mechanisms that enterprises require.
OpenAI also views security as a crucial prerequisite for broader AI adoption. CISO Dane Stuckey states that the collaboration aims to help organizations identify risks more quickly, increase their digital resilience, and use AI in production environments in a controlled and compliant manner.
Increasing Focus on AI Security
Through this collaboration, IBM gains access to the specialized cybersecurity capabilities that OpenAI has recently developed. For OpenAI, the partnership represents a new step toward making its models available not only for generative AI applications but also for security purposes within large enterprises.
The new application security service is available immediately. IBM says that additional integrations within the OpenAI Daybreak Cyber Partner Program will follow in the coming period.