OpenAI: AI Researcher by 2028

OpenAI CEO Sam Altman has announced an ambitious goal to develop a fully autonomous AI researcher by 2028. This AI would be capable of independently tackling major scientific breakthroughs. As a stepping stone, OpenAI aims to have an intern-level AI research assistant by September 2026, capable of data analysis and simple experiments.

This initiative relies on algorithmic innovation and significantly scaling up 'test time compute'. OpenAI's corporate restructuring is set to enable the $1.4 trillion infrastructure investment required to achieve these goals. The company is working towards systems that can accelerate scientific and technological discovery, compressing years of innovation into months. Current AI models already match top humans in math competitions and handle five-hour task horizons, paving the way for rapid advancement.

OpenAI also emphasises the importance of secure and transparent AI systems aligned with human values. This includes ensuring user privacy and preventing the misuse of AI for harmful purposes, such as engineering biological threats. The company is actively studying ways to detect and prevent such misuse, including building secure model access systems and enhancing monitoring around sensitive capabilities.