Morgan Stanley built its in-house AI tool to tackle a difficult coding problem: reworking old legacy code into more updated coding languages. Morgan Stanley introduced the AI tool, which is based on OpenAI’s GPT models, in January. The tool, called DevGen.AI, translates code in older languages, such as Perl into plain English, which developers can then use as a basis for rewriting the code into newer languages like Python. Mike Pizzi, Morgan Stanley’s global head of technology and operations, told WSJ that in the five months since its launch, DevGen.AI has worked through nine million lines of code, saving the firm’s 15,000 developers roughly 280,000 hours of work. Pizzi said that Morgan Stanley opted to build the tool itself because tech companies didn’t have any solutions that could fit Morgan Stanley’s exact specifications. Commercial tools lacked expertise in deciphering older coding languages, especially those specific to a company. Morgan Stanley trained DevGen.AI on languages within its own code base, including languages customized for the company. However, the AI tool still has growing to do when it comes to full translation. Though the tool can, in theory, rewrite code from an older language to a newer one, it doesn’t know how to write the new code efficiently or as well as a human developer, Pizzi said. That’s why Morgan Stanley is keeping human developers involved in the process of translating old or legacy code to new languages. Pizzi disclosed that the firm will not be reducing its software engineering workforce as a result of the AI tool.