OpenAI president Greg Brockman says AI now writes up to 80 percent of code as software development rapidly shifts toward human oversight

Brockman’s remarks highlight how quickly AI driven coding systems have evolved. He explained that as recently as December, these tools were generating roughly 20 percent of code within projects. In just a short span of time, that share has surged dramatically.

This change reflects not only improvements in AI capabilities but also growing trust among developers and organizations. Instead of using AI for small snippets or repetitive tasks, teams are increasingly relying on it for substantial portions of application logic, infrastructure setup, and testing.

Brockman noted that this transition fundamentally changes the developer’s role. AI is no longer a side feature in the development process. It is becoming the main engine behind it.

The rise of AI generated code is fueling a wider debate across the technology industry. At the center of this discussion is a simple but profound question: are developers becoming supervisors rather than primary creators?

Brockman acknowledged this shift while emphasizing that human responsibility remains central. At OpenAI, every line of code that enters production is still reviewed and approved by a human engineer. This ensures accountability, quality control, and alignment with broader system goals.

Rather than replacing developers, AI appears to be reshaping their responsibilities. Engineers are increasingly focusing on reviewing outputs, refining prompts, designing system architecture, and ensuring reliability. The craft is evolving from writing every line manually to guiding intelligent systems effectively.

Brockman also pointed to the growing accessibility of AI coding tools, particularly platforms like . Initially designed with professional developers in mind, such tools are now expanding their reach to a much broader audience.

Today, individuals with minimal programming experience can generate functional code, build prototypes, and automate workflows using AI assistance. This democratization of software creation is expected to lower barriers to entry and enable more people to participate in digital innovation.

However, this wider access also raises questions about code quality, security, and long term maintainability, especially when non experts rely heavily on automated systems.

OpenAI’s experience is not an isolated case. Several major technology companies are witnessing similar trends, reinforcing the idea that AI driven coding is becoming an industry standard.

, chief executive of , has stated that around 75 percent of new code written within the company is now generated by AI systems before being reviewed by engineers.

At , internal expectations suggest that a large share of engineers, particularly those working on core product experiences, will soon rely on AI for the majority of the code they commit. Reports indicate that this could involve AI generating more than three quarters of their contributions.

Meanwhile, of has offered even more ambitious projections. He suggested that AI could write up to 90 percent of code within a matter of months, and potentially nearly all code within a year.

Taken together, these perspectives point to a clear direction: AI is not just augmenting software development, it is redefining it.

Despite the rapid adoption, Brockman cautioned against extreme approaches. He emphasized that neither blind reliance on AI nor outright rejection is the right path forward.

Instead, he advocated for a balanced strategy where AI is used thoughtfully and responsibly. This includes maintaining rigorous review processes, ensuring accountability, and understanding the limitations of automated systems.

The importance of human judgment becomes even more critical as AI generated code scales. Issues such as hidden bugs, security vulnerabilities, and unintended behavior can arise if outputs are accepted without scrutiny.

This balance reflects a broader principle that has guided technological progress over the years. New tools may change how work is done, but human oversight remains essential for ensuring quality and trust.

The shift toward AI generated code carries far reaching implications for the future of the software industry. For developers, it means adapting to new workflows and acquiring new skills centered around collaboration with AI systems.

Prompt engineering, system design, and critical evaluation of AI outputs are likely to become core competencies. Traditional coding skills will still matter, but they may play a different role within a more automated environment.

For companies, the productivity gains could be significant. Faster development cycles, reduced costs, and the ability to scale projects more efficiently are among the potential benefits. At the same time, organizations will need to invest in governance frameworks to manage risks associated with AI generated code.

For the broader ecosystem, the barrier to creating software may continue to fall. This could lead to an explosion of innovation, but also an increase in low quality or insecure applications if proper standards are not maintained.

In many ways, the current moment resembles earlier turning points in computing history. The move from assembly language to high level programming languages, and later to frameworks and automation tools, each redefined how developers worked.

AI represents the next step in that progression. Instead of simply abstracting complexity, it actively participates in the creative process of building software.

Brockman’s remarks suggest that this transition is happening faster than many anticipated. What once seemed like a distant possibility is now becoming an everyday reality in engineering teams.

Even as AI takes on a larger share of coding tasks, the principle of accountability remains unchanged. Brockman made it clear that at OpenAI, a human is always responsible for the final output.

This approach reflects an understanding that technology, no matter how advanced, operates within human defined systems. Decisions about what code is deployed, how it is used, and its broader impact ultimately rest with people.

Maintaining this responsibility will be crucial as AI systems continue to evolve and take on more complex roles.

The trajectory outlined by industry leaders points toward a future where AI generated code becomes the norm rather than the exception. Whether that future arrives in months or years, the direction is increasingly clear.

For developers, the challenge will be to adapt without losing the discipline and rigor that define good engineering. For companies, it will be about harnessing AI’s potential while safeguarding quality and security.

And for the industry as a whole, the task will be to ensure that this powerful shift benefits not just productivity, but also reliability, trust, and long term innovation.

Brockman’s observations offer a glimpse into this unfolding transformation. As AI continues to reshape software development, one thing remains certain: the relationship between humans and machines is entering a new phase, one where collaboration, rather than replacement, will define success.