Prafulla Dhariwal: Pune’s genius behind GPT-4o's Ghibli-inspired AI art

The GPT-4o image generation model has become a cultural phenomenon, especially in India, with users embracing the "Ghiblification" trend, blending creativity and technology, while Dhariwal sees it as a testament to AI's power in artistic expression

Prafulla Dhariwal, the innovative mind behind GPT-4o’s image-generation capabilities, has quickly become a household name, especially in India, where his AI-powered creation has sparked a creative revolution. The 30-year-old Head of Multimodal at OpenAI has witnessed his brainchild, the GPT-4o image generator, become an instant sensation, particularly for its ability to transform photographs and images into Ghibli-style art.

Dhariwal, a Pune native, credits the overwhelming response to the image generation model, particularly from India, as one of the most rewarding aspects of his career. "We knew we were building something meaningful, but the scale of the response—especially from India—completely blew us away," he said in a recent interview. The model has gained such widespread popularity that India is now OpenAI's fastest-growing market for image generation, a milestone Dhariwal calls the "biggest growth surge in OpenAI’s history."

Born and raised in Pune, Dhariwal's journey to the forefront of AI innovation is one of perseverance and ambition. He attended Dr. Kalmadi Shamrao High School before pursuing higher education at the prestigious Massachusetts Institute of Technology (MIT), where he graduated in 2017. His journey in AI began with an internship at OpenAI in 2016, which later led to a full-time role at the company. He attributes his success to persistence, humility, and a clear sense of purpose.

GPT-4o transforms creativity with Ghibli art

The GPT-4o image generation model has become a cultural phenomenon, particularly in India, where users have embraced the "Ghiblification" trend. From selfies to iconic photographs, the model has enabled users to create images in the beloved Ghibli style, blending creativity with technology. Dhariwal views this development as a testament to the power of AI in enabling artistic expression, despite some controversies around copyright and ethics.

OpenAI CEO Sam Altman has lauded Dhariwal for his vision and determination, stating that GPT-4o would not have been possible without him. "The launch of image generation in GPT-4o has been one of the most rewarding milestones in my time at OpenAI," Dhariwal reflects, calling it both technically ambitious and creatively energizing.

While visiting India for his wedding preparations last December, Dhariwal used the model to create design ideas for wedding invitations. He recalls, “Seeing something I had built become part of such a meaningful moment was truly special.”

The development of the image generator took nearly a year, with Dhariwal and his team working tirelessly to make the system faster, sharper, and more intuitive. OpenAI’s Chief Operating Officer, Brad Lightcap, described the first week of image generation as "very crazy," with over 130 million users generating more than 700 million images. The tool has proven incredibly popular, especially in India, where creative individuals have used it to generate everything from family portraits to Bollywood-inspired scenes.

Dhariwal talks AI ethics and future impact

Addressing concerns around ethics and copyright, Dhariwal emphasized that OpenAI's policies are designed to respect both artistic freedom and legal boundaries. "We don't allow generations in the style of individual living artists, but we permit broader studio-inspired styles, as style itself isn't copyright-protected," he explained. The company continues to adapt its policies as real-world use evolves, ensuring that users' creativity remains both meaningful and respectful of intellectual property.

Looking ahead, Dhariwal is excited about the potential for collaborations with animation studios and artists, as well as exploring new ways to create authentic, licensed experiences. "The creativity coming out of India has been incredible," he says, mentioning that users have even created designs for wedding invitations and planned proposals using the AI tool.

As a cricket fan, Dhariwal was particularly humbled when Indian cricket legend Sachin Tendulkar shared an image generated with GPT-4o. He believes this moment encapsulated the global resonance of the tool.

Dhariwal encourages India’s AI talent

Dhariwal’s message to aspiring AI researchers and students, particularly in India, is clear: "Stay curious, keep learning, and embrace hard problems. AI is moving fast, and there’s room for people from all backgrounds to contribute." His vision for the future includes tools that empower people from all walks of life, whether they're students, teachers, designers, or engineers.

In ten years, Dhariwal hopes to have helped create AI tools that facilitate learning, boost productivity, and enable self-expression in ways previously unimaginable. With India's immense untapped creativity, he sees boundless opportunities for impact and fulfilment. "If we can help someone work more efficiently, learn more intuitively, or express themselves more effectively—that will be deeply fulfilling," he concludes.

Also Read: Sam Altman Hints at ‘Images v2’ as ChatGPT Faces Surge in Ghibli-Style Art Demand