AI’s creative leap by 2026: skepticism grows over true scientific breakthroughs

OpenAI CEO Sam Altman forecasts that by 2026, artificial intelligence will transcend data analysis to generate novel scientific insights, marking a significant creative leap.

Altman envisions AI reshaping work, energy, innovation, and scientific discovery as part of humanity's extended evolution. Supporting this vision, OpenAI President Greg Brockman highlighted the April 2025 release of the o3 and o4-mini models, which "were the first models scientists used to create useful new ideas in their field."

Despite this optimism, skepticism persists among AI experts. Thomas Wolf, Chief Science Officer at Hugging Face, cautions that current AI models lack the ability to pose genuinely new questions, a critical step for true scientific breakthroughs.

"These were the first models scientists used to create useful new ideas in their field," Brockman said, underscoring OpenAI's progress.

"Existing models can't pose genuinely new questions, a precondition for any scientific breakthrough," Wolf argued, reflecting ongoing doubts.

This debate highlights the tension between AI's rapid advancements and the challenges of achieving authentic creativity and innovation. While OpenAI pushes forward with promising models, the scientific community remains cautious about AI's capacity to independently drive groundbreaking discoveries.

As 2026 approaches, the AI field stands at a crossroads between hopeful anticipation and critical scrutiny regarding its creative potential.

OpenAI's Sam Altman predicts AI will achieve novel scientific insights by 2026, signaling a leap from data analysis to creative thinking. However, skepticism grows among experts like Hugging Face's Thomas Wolf, who doubts current AI models can generate truly new scientific questions, essential for breakthroughs.