AI could soon be cheaper than human labor: Sam Altman
What's the story
OpenAI CEO Sam Altman has warned that artificial intelligence (AI) is on track to become cheaper than human labor. In an interview with Forbes, he explained that the cost of operating AI models at the 'inference' stage, when they produce outputs, is already much lower than the energy needed for humans to do similar intellectual work.
Cost comparison
AI's efficiency and future improvements
Altman challenged the usual comparisons that stack AI's training costs against human effort. He noted that human intelligence takes decades of biological 'training' and energy consumption over a lifetime. On a per-unit basis of intellectual output, he said, AI systems are becoming highly efficient and this efficiency is only going to improve further.
Energy concerns
The impact of AI on jobs
Altman also warned that despite AI being cheaper per task, its widespread use could increase total global energy consumption. He said the technology will have a 'big impact' on jobs, but not uniformly. Some roles will be partially automated while others may be completely transformed with new categories of work emerging, similar to past technological shifts.
Skill evolution
Skills for the future
Altman stressed the importance of adaptability, resilience, and fluency in using AI tools over narrowly focusing on coding as a universal skill. He also highlighted resource allocation as an important aspect of AI's growth. A growing share of resources is now going toward 'inference,' or the continued use of AI, rather than initial model training.
Market dynamics
India's dynamic AI market
Altman described India as one of the most dynamic markets for AI adoption, saying the country is investing across the full stack from infrastructure to applications. He said India is already OpenAI's fastest-growing market for its coding product and could soon become its largest. However, he warned that delivering AI services costs more than traditional internet services, so companies will have to make such markets financially viable at scale.
Educational shift
Adapting education systems to tech shifts
On fears of AI making humans think less, Altman drew parallels with past tech shifts like search engines. He said the real risk isn't the tech itself but whether education systems adapt. If teaching methods don't change, students could become overly reliant on AI tools. Altman also pointed to safety and privacy concerns as major roadblocks in adopting AI-powered personal agents widely.