On Monday, OpenAI released Dawn, a cybersecurity initiative designed to use artificial intelligence to help developers and security teams identify vulnerabilities, verify fixes, and protect software more quickly.
This announcement highlights that AI companies are increasingly moving into the cybersecurity domain, as advanced models continue to improve in analyzing code, identifying software vulnerabilities, and automating technical tasks, signaling an expansion of the transformation within the AI field.
At the Postal X conference, OpenAI CEO Sam Altman referred to Daybreak as "an effort to accelerate network defense and continuous software protection."
“Artificial intelligence is already excellent at cybersecurity and is about to become even better; we want to start partnering with as many companies as possible now to help them continuously protect their security,” Altman wrote.
OpenAI stated that Daybreak combines the company’s AI models with its coding-focused agent system, Codex, to help security teams review code, analyze dependencies, simulate threats, validate patches, and investigate unfamiliar systems. The company said its goal is to reduce the time between discovering and fixing vulnerabilities.
OpenAI has not yet responded to this matter.Decrypt.
Amid warnings from cybersecurity researchers and industry experts about the threat of AI-driven cyberattacks, the magazine Dawn has emerged. Last month, Mozilla, the developer of the Firefox browser, stated that they used Mythos to identify 271 previously unknown vulnerabilities in the browser.
OpenAI stated in a statement: “Artificial intelligence can now help defenders reason across codebases, identify subtle vulnerabilities, validate fixes, analyze unfamiliar systems, and transition more quickly from discovering vulnerabilities to patching them. Because these capabilities can also be misused, Daybreak combines enhanced defensive capabilities with trust, verification, appropriate security safeguards, and accountability.”
Meanwhile, major AI companies are increasingly deploying their models into cybersecurity and software engineering. Anthropic, a competitor of OpenAI, is also placing greater emphasis on AI-related business initiatives. Publicly traded As competition among AI companies intensifies in their pursuit of enterprise clients, its Claude model can be used for coding and security-related tasks.
Experts still split on the extent of the threat posed by AI, as researchers and government agencies have warned that advanced AI models could accelerate this threat by helping hackers automate vulnerability research, malware development, and exploit creation. Meanwhile, Google researchers recently stated that large language models are becoming increasingly adept at identifying and exploiting software vulnerabilities that traditional security scanners often overlook.
OpenAI stated that it plans to collaborate with government and industry partners before deploying more AI models with cybersecurity capabilities, as regulators and national security officials are attempting to closely review testing of advanced AI models before their public release.
OpenAI writes: "Dawn is the first light of morning. For cybersecurity, this means identifying risks earlier, acting sooner, and helping software be resilient from the ground up."