Some of the world’s most important institutions still depend on decades-old software written in languages like COBOL. These systems quietly support banking, tax processing, airline operations, and government infrastructure — but maintaining them has become increasingly difficult as experienced programmers retire.
OpenAI says its newest models show substantial improvements in understanding and migrating older programming languages into modern code. That progress could give organizations a powerful new way to analyze legacy systems, generate safer migration plans, and reduce the risk of touching code that has long been considered too fragile to change.
Why this matters
- Critical infrastructure: Legacy systems are deeply embedded in essential public and private services.
- Talent gap: Fewer developers are trained to work with older languages like COBOL.
- Modernization boost: AI could help translate, document, and test old code more efficiently.
While these tools are not a magic switch — especially for high-stakes systems where accuracy and reliability are crucial — they represent a meaningful step toward solving a long-running technology challenge. If AI can help organizations safely modernize legacy software, the benefits could be felt across industries and public services worldwide.