BETA RELEASE

Summary

A scientific assessment of general-purpose AI capabilities, emerging risks (misuse, malfunction, systemic), and risk management approaches to support informed policymaking.

Key quotes

General-purpose AI systems can perform a wide range of well-scoped tasks with high proficiency
The capabilities of these systems are ‘jagged’: they can perform many complex tasks but still struggle with some seemingly simpler ones.
If current trends continue, AI systems could operate autonomously on multi-day tasks by 2030

This document provides an extended summary of the 2026 International AI Safety Report, detailing progress in AI agents and reasoning models. It analyzes risks including biological weapon misuse and labor market shifts.