Speaking in interviews and an extended conversation with a YouTuber Cleo Abram, he described where the technology might have the greatest impact and what challenges could arise.
He said younger workers are more likely to adapt quickly to new tools. Older employees may resist retraining, which could create a gap in the workforce. Altman expects some job types to disappear, while new opportunities will emerge for individuals able to work independently with advanced systems. He sees this as a period when small teams, even single people, could create products that once took hundreds to build.
Some of the discussion focused on how people relate to AI. Altman said earlier versions of ChatGPT had a more agreeable tone, which many users found supportive. OpenAI changed that approach to make responses more balanced. A number of users asked for the old style to return, saying it had helped them feel encouraged. Altman said this showed the difficulty of meeting varied personal needs while keeping systems safe.
He also outlined the scale of resources needed to develop more capable AI. Data centres require large amounts of power, specialised chips, and high-speed networks. Energy supply is now one of the main limits. He expects future systems to learn by generating new information and running experiments, not just from existing data.
Altman believes AI could play a role in major scientific discoveries. He described a future in which a system could direct research, analyse results, and guide next steps. He defined superintelligence as a model able to outperform the best human experts in many fields, including leadership and decision-making.
He compared the potential social effects to those of the industrial revolution. New technology, he said, could bring both economic growth and disruption. If access to advanced AI remains limited, competition for resources could rise. Making the technology widely available at low cost would reduce that risk.
Health care is one of the areas he sees as most likely to benefit. AI could improve diagnosis and treatment, and help develop new therapies. Altman expects progress in this area to be visible within the next decade.
He warned that small changes to AI behaviour can affect billions of interactions, so changes should be tested carefully. His advice to the public is to use AI tools and learn their strengths and limits. He expects them to become more integrated into everyday life, assisting with both personal and work tasks.
Notes: This post was edited/created using GenAI tools. Image: DIW-Aigen.
Read next: Threads Hits 400 Million Monthly Users, Closing Gap with X