On 2026-09-03, US Senator Bernie Sanders and Representative Greg Casar announced a proposal for the "Ban Artificial Superintelligence Act," which aims to permanently prohibit the development and deployment of Artificial Superintelligence (ASI). As of that day, the full text and bill number had not been released, and the proposal remained at the stage where the proponents provided an outline.

The proposal targets systems that match or exceed human cognitive abilities across a wide range of domains, or systems capable of executing plans that could strip humanity of its authority. Furthermore, it indicates a policy to pause the development of advanced AI until a new federal regulatory agency establishes safety rules and model review procedures. The proposed framework includes the creation of a cabinet-level federal agency to monitor frontier AI and oversee the removal of dangerous capabilities or the avoidance of shutdown orders, as well as sanctions such as prison terms of up to 20 years for violators.

On the same day, OpenAI announced its new AI model, "GPT-6 Astra." The company positions Astra as its most high-performance model to date, with high capabilities in areas such as computer operation and cybersecurity. According to OpenAI's safety evaluations, Astra's cybersecurity capabilities reached the "Critical" level for the first time. In a briefing for reporters, OpenAI President Greg Brockman expressed the view that it would not be unreasonable to consider the present as the era of AGI (Artificial General Intelligence).

Source: