Prospects of LLMs Leading to Artificial General Intelligence (AGI)

Determine whether generative artificial intelligence systems, particularly large language models trained via next-token prediction on massive text corpora, will lead to artificial general intelligence (AGI).

Background

In discussing the timing and trajectory of statistical engagement with LLMs, the paper notes a key uncertainty regarding the long-term outcome of generative AI. Clarifying whether LLMs can lead to AGI has broad implications for research priorities, societal impact, and the urgency of establishing statistical foundations.

Regardless of the outcome, the author argues that generative AI technologies will likely remain a significant component of the AI landscape, underscoring the importance of timely statistical contributions.

References

It remains uncertain whether generative AI, particularly LLMs, will lead to artificial general intelligence.

— Do Large Language Models (Really) Need Statistical Foundations?  (2505.19145 - Su, 25 May 2025) in Section “Discussion,” final paragraph

We defend, further, a deliberately strong conjecture: LLMs cannot scale to superhuman intelligence while knowledge remains organised as statistical association between substrings.

— Semantic Bayesian World Models  (2609.03834 - Soru, 3 Sep 2026) in Section 1, Introduction; Section 3.4, Why Symbols Matter