Safe Code Warrior launches AI governance studying


Joseph Gabriel Lagonsin


JOSEPH GABRIEL LAGONSIN

Information Editor

Safe Code Warrior has launched Adaptive Studying for its AI software program governance platform. The function is meant to attach developer coaching to real-time coding exercise.

It hyperlinks the corporate’s Belief Agent product with its studying platform and triggers coaching when builders use AI instruments or when vulnerabilities are recognized in code repositories. The purpose is to provide organisations proof of coaching outcomes on the commit degree.

The launch comes as corporations face rising considerations about how AI is reshaping software program growth practices and safety oversight. Safe Code Warrior pointed to a pointy improve in code churn linked to AI adoption, citing Faros analysis exhibiting that the ratio of traces deleted to traces added in merged code rose 861% every quarter.

The corporate additionally cited trade knowledge exhibiting wider use of AI instruments by workers on firm units, alongside rising use of non-corporate accounts to entry these providers. Based on the figures it referenced, supply code is now the commonest knowledge kind submitted to unauthorised exterior AI fashions, rising the chance of mental property publicity.

Vulnerability administration is one other strain level. Safe Code Warrior mentioned vulnerability exploitation has overtaken credential abuse because the main breach methodology, whereas solely 26% of crucial vulnerabilities had been totally remediated in 2025 and median remediation time elevated to 43 days.

The way it works

Adaptive Studying is constructed round two parts: AI Indicators and Vulnerability Indicators. AI Indicators are designed to detect which AI instruments particular person builders are utilizing and join that exercise to focused coaching primarily based on the code they commit.

Vulnerability Indicators join knowledge from safety instruments to studying assignments. The function can establish vulnerabilities in repositories the place builders are working and assign coaching related to these points and the code underneath growth.

It additionally consists of coverage controls that allow prospects outline when coaching ought to be triggered primarily based on AI utilization indicators. The system can establish builders interacting with AI instruments by noticed exercise, assign coaching duties by the corporate’s quest format, and monitor completion and efficiency over time.

On the vulnerability facet, the system helps software programming interface knowledge imports from Checkmarx, SonarQube and Parasoft, alongside SARIF uploads. Scheduled synchronisation is deliberate.

Governance focus

The launch expands Safe Code Warrior’s broader push into AI software program governance, an space that has gained urgency as companies search to know how AI-generated and AI-assisted code is created, reviewed and deployed. The corporate mentioned its Belief Agent product was constructed to make AI affect in software program growth seen, attributable and enforceable.

Adaptive Studying is meant to maneuver governance from monitoring to intervention by tying coaching on to dangerous behaviour and code modifications. It additionally creates auditable data of per-developer AI safety coaching linked to manufacturing code.

Based on Safe Code Warrior, these data can help compliance efforts tied to the EU AI Act, ISO/IEC 42001 and the NIST AI Threat Administration Framework, positioning the product in a market the place software program governance instruments are more and more a part of broader compliance and threat administration programmes.

Pieter Danhieux, Co-founder and Chief Govt Officer of Safe Code Warrior, outlined the corporate’s view of what enterprises now want from AI oversight in software program engineering.

“At each stage, enterprises try to realize three major targets: builders and brokers should be taught to construct securely, companies should govern what AI can and might’t contact within the codebase, and safety groups should be capable of hint which AI did what, the place, and for whom,” mentioned Pieter Danhieux, Co-founder and Chief Govt Officer of Safe Code Warrior.

He mentioned the brand new function is designed to show that visibility into motion.

“With SCW’s Adaptive Studying, organisations and builders can swiftly transfer from understanding threat, to actively lowering it at scale, and with measurable proof on the commit degree. That is crucial as builders transfer from extra conventional workflows, to environments the place they’re orchestrators of autonomous brokers,” Danhieux mentioned.