His comments come at a time when world leaders are debating how much control they should have over AI.
Six months into 2026, the AI landscape is advancing at an unprecedented pace. Key developments include the launch of several major frontier models, a new round of funding for Anthropic, a revised contract with OpenAI and Microsoft, and expanded integration of AI into defense operations.
These developments signal a shift in the way the world perceives and uses AI. No longer limited to chatbots and shopping recommendations, AI is now reshaping manufacturing, healthcare, finance, national security, and more.
At the G7 Summit, OpenAI CEO Sam Altman urged world leaders not to rely solely on AI companies and research institutes to build AI governance. Instead, it called on governments to set global standards for implementing rapidly advancing technology.
“Don’t hand over your responsibility to an AI lab like me,” Altman told leaders and technology executives.
His comments come at a time when world leaders are debating how much control they should have over AI.
“We develop the technology and the people of the free world make the rules,” Altman said. “Engineers have special knowledge about AI, but they don’t have special wisdom about humanity.”
The head of OpenAI says the question of whether AI will be useful is “settled” and he expects to see a system of “amazing power” within the next year or two that can reshape humanity.
Among all sectors, AI Lab has prioritized healthcare and medicine as use cases.
More than 230 million people turn to ChatGPT for health and wellness advice each week, so earlier this year the company launched ChatGPT Health, a dedicated workspace within the platform designed to securely analyze personal medical records and wellness data.
OpenAI researcher Karan Singhal co-leads a team working on biomedical AI, fundamental models, AI safety, and representation learning that aims to revolutionize the way people use and trust models in OpenAI. The goal is for patients to trust this tool as their “guardian in the treatment process.”
The GPT-5 model family is a step in that direction.
“We definitely want models to be ahead of everything else,” Singhal told the media.
Recently, the model helped doctors diagnose rare genetic diseases affecting children. Using OpenAI’s o3 Deep Research inference model, researchers from Boston Children’s Hospital Menton Orphan Disease Research Center, Harvard University, and OpenAI examined anonymized clinical and genomic data extracted from 376 cold cases. They established the diagnosis in 18 cases.
“The bottleneck is time. Specialists can only spend so much of their day working on a particular person,” said Dr. Katherine Brownstein of Boston Children’s Hospital’s Menton Center for Orphan Disease Research.
Healthcare remains one of the most important and fastest growing frontiers for AI, with the global AI in healthcare market expected to reach USD 194.79 billion by 2031, up from USD 36.67 billion in 2026.
