AI Reverse Engineering Replicates Software, Sparks Controversy

Brandon Thomas, a U.S. software developer, recently created software with similar functionalities to seven programs, including Adobe’s Photoshop, Illustrator, and Premiere Pro, using artificial intelligence (AI). He later developed programs mimicking Microsoft’s Word, Excel, and PowerPoint. The founder of AI video and voice generation company Storyteller AI, Thomas publicly released these programs for free. As news spread that core features of existing commercial software could be replicated with AI, financial markets reacted. British financial firm HSBC lowered Adobe’s investment rating from “Buy” to “Hold” on the 13th and reduced its target stock price from $310 to $267, a 13.9% decrease.
Cases of replicating Adobe and Microsoft programs have surged, making AI-based “reverse engineering” a new hot topic in the software industry. Reverse engineering involves analyzing the internal structure and operation of completed software to implement similar functions in new code. It is akin to disassembling a newly released car to understand its engine and transmission, then using that knowledge to build a similar vehicle.
◇Reporter’s Hands-On Test…Document Editing Function Recreated in 3 Minutes
While reverse engineering software has existed, its recent proliferation accelerated after the release of the AI reverse engineering tool “REA (Reverse Engineer Anything)” on GitHub, the world’s largest software development platform. REA connects AI coding tools like OpenAI’s Codex and Anthropic’s Claude Code with specialized reverse engineering tools to analyze a program’s structure and operation. Notably, it allows users to input only necessary information from vast analysis results, reducing token (data processing unit) usage and costs.

A reporter tested REA and OpenAI Codex to recreate parts of “Novel,” an open-source document editor. Investigating the program’s structure with REA took approximately 7 seconds. Codex then read the open-source code, newly implemented and verified editing functions like titles, bold text, and bullet points in 3 minutes and 3 seconds. All 15 tests comparing the original and the result matched. An additional 3 minutes and 44 seconds were spent creating the editing interface, which passed 14 operational tests. When the reporter input Korean text and applied titles and bold formatting, it functioned normally.
AI reverse engineering is gaining attention because it can analyze not only existing software but also the design methods and development know-how accumulated by companies over years. Unlike “vibe coding,” where users describe desired functions to create programs, reverse engineering identifies implementation methods by analyzing completed programs.
Industry reactions to AI reverse engineering are mixed. Dutch developer and entrepreneur Peter Levels stated, “Software is dead. People will create programs themselves or ask AI to do so whenever needed.” Brazilian developer Fábio Akita countered, “AI-generated alternatives have errors and missing features, and it’s difficult to analyze all functions of commercial products externally.”

◇Domestic Attempt to Recreate ‘HWP’…Copyright Dispute Risks
A South Korean developer publicly revealed “Tangul2026,” a program that reads and edits Hancom’s HWP and HWPX files using AI reverse engineering tools. The developer explained it was created for educational and demonstration purposes. This marks the first domestic attempt to recreate commercial software functions using AI.
As AI-based reverse engineering and function replication attempts increase globally, legal disputes over copyright infringement may rise. South Korea’s Copyright Act permits analyzing a program’s operational principles or code for compatibility under certain conditions. However, using obtained information beyond permitted purposes or unauthorized replication of protected code from the original can lead to copyright violations.
☞Software Reverse Engineering
A technology that uses AI to understand a software’s internal structure and operation, then implements it in new code. It involves decompiling executable files into code and observing responses to various inputs (commands). AI coding tools create new programs based on this analysis and compare/verify if they function like the original.



