Will at least one AI system publicly demonstrated before January 1, 2030 meet the criteria for Artificial General Intelligence (AGI)?

RESOLUTION CRITERIA

This market resolves YES if, before January 1, 2030, a single AI system is publicly demonstrated that satisfies ALL of the following conditions:

1. Performs at or above human expert level on a broad range of cognitively demanding tasks spanning at least 5 distinct domains (e.g. mathematics, coding, science, writing, reasoning, medicine).

2. Can learn and adapt to entirely new tasks with minimal instruction (few-shot or zero-shot), without task-specific retraining.

3. Can autonomously plan and execute long-horizon tasks (multi-day, multi-step agentic work) that generate measurable real-world value.

4. Is widely recognized as AGI by credible AI researchers or major AI labs (e.g. OpenAI, DeepMind, Anthropic, Meta AI), OR meets OpenAI's own stated AGI threshold: "highly autonomous systems that outperform humans at most economically valuable work."

Note: Narrow AI advances (e.g. AlphaFold, GPT-class models excelling in a single domain, or AI that wins at specific games) do NOT qualify. The system must demonstrate general cross-domain capability comparable to a skilled human generalist.

BACKGROUND

AGI (Artificial General Intelligence) refers to a hypothetical AI that can match or surpass human cognitive abilities across virtually all tasks — unlike today's narrow AI systems, which excel in specific domains. Creating AGI is a stated mission of major AI labs including OpenAI, Google DeepMind, and Anthropic.

WHAT EXPERTS SAY

• 2022 AI Impacts Expert Survey (738 ML researchers): Median estimate for 50% probability of human-level machine intelligence was ~2059. Only ~10% of respondents put the probability of AGI before 2030 at more than 25%.

• Metaculus community forecast (1,800+ forecasters, as of early 2026): Median estimate for first general AI system is October 2032 — just two years after our deadline. https://www.metaculus.com/questions/5121/

• Sam Altman (OpenAI CEO, 2024): Suggested AGI could arrive "in the next few years." OpenAI's internal roadmap reportedly targets AGI in the late 2020s.

• Demis Hassabis (Google DeepMind CEO): Has stated DeepMind is pursuing AGI as a long-term goal; cautiously optimistic about timelines within the next decade.

KEY CONSIDERATIONS

Arguments FOR (YES): Rapid scaling of LLMs, emergence of capable AI agents, trillion-dollar investment in AI infrastructure, competitive pressure among major labs, and improvements in reasoning/planning suggest the pace of progress could surprise even optimists.

Arguments AGAINST (NO): Current AI systems still struggle with robust reasoning, grounding, planning under uncertainty, and common sense. Scaling laws may hit diminishing returns. Regulatory scrutiny, compute constraints, and fundamental architectural limitations could slow progress. Most independent researchers consider 2030 an aggressive timeline.

RESOLVER NOTES

The creator will resolve based on credible, peer-reviewed reports or major announcements from leading AI laboratories, with corroboration from multiple independent expert sources. The bar is high: marketing claims alone (e.g., a company declaring its own product "AGI") will not suffice without substantiated third-party verification. If AGI is genuinely ambiguous at the deadline, this will resolve N/A and all positions will be refunded.