Stop treating `AGI' as the north-star goal of AI research

2025/02/06 by Borhane Blili-Hamelin, Christopher Graziul, Blili-Hamelin, Borhane +30 · 20 voices · 13 citations
Computer Science · Medicine · Psychology · #Artificial Intelligence in Healthcare and Education #Astrophysics #Physics #Psychology #Star (game theory) #cs.CY

paper · pdf · doi:10.48550/arxiv.2502.03689

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2025/02/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

The AI research community plays a vital role in shaping the scientific, engineering, and societal goals of AI research. In this position paper, we argue that focusing on the highly contested topic of `artificial general intelligence' (`AGI') undermines our ability to choose effective goals. We identify six key traps -- obstacles to productive goal setting -- that are aggravated by AGI discourse: Illusion of Consensus, Supercharging Bad Science, Presuming Value-Neutrality, Goal Lottery, Generality Debt, and Normalized Exclusion. To avoid these traps, we argue that the AI research community needs to (1) prioritize specificity in engineering and societal goals, (2) center pluralism about multiple worthwhile approaches to multiple valuable goals, and (3) foster innovation through greater inclusion of disciplines and communities. Therefore, the AI research community needs to stop treating `AGI' as the north-star goal of AI research.

Citations

• A metatheory of classical and modern connectionism.
• Reclaiming AI as a Theoretical Tool for Cognitive Science
• Beyond Preferences in AI Alignment
• Position: Why We Must Rethink Empirical Research in Machine Learning
• A Roadmap to Pluralistic Alignment
• Levels of AGI for Operationalizing Progress on the Path to AGI
• Language Models Represent Space and Time
• Making Intelligence: Ethical Values in IQ and ML Benchmarks
• The Benchmark Lottery
• On the Dangers of Stochastic Parrots
• Towards Decolonising Computational Sciences
• The Hardware Lottery
• Fairness and Abstraction in Sociotechnical Systems
• Minds, brains, and programs

Cited by

• Legal Alignment for Safe and Ethical AI
• AI for scientific discovery is a social problem
• Deep Hype in Artificial General Intelligence: Uncertainty, Sociotechnical Fictions and the Governance of AI Futures
• What the HellaSwag? On the Validity of Common-Sense Reasoning Benchmarks
• Epistemic diversity across language models mitigates knowledge collapse
• Mapping Data Labour Supply Chain in Africa in an Era of Digital Apartheid: a Struggle for Recognition
• On the Measure of a Model: From Intelligence to Generality
• What if AI systems weren't chatbots?
• Making Power Explicable in AI: Analyzing, Understanding, and Redirecting Power to Operationalize Ethics in AI Technical Practice
• World(coin) in the AI future: how sociotechnical fictions are instrumental to the cyberlibertarian transition
• Is the `Agent' Paradigm a Limiting Framework for Next-Generation Intelligent Systems?
• NLP Meets the World: Toward Improving Conversations With the Public About Natural Language Processing Research
• Rigor in AI: Doing Rigorous AI Work Requires a Broader, Responsible AI-Informed Conception of Rigor

Discussions

• Our new work deconstructing the concept of "AGI", led by the brilliant @borhane.bsky.social in collaboration with @graziul.bsky.social , Hananel Hazan, @shiridoshi.bsky.social +more co-authors below, [bsky, 90 points, 1 comments]
• Coming to an ICML near you arxiv.org/abs/2502.03689 [bsky, 58 points, 1 comments]
• Stop treating `AGI' as the north-star goal of AI research [hn, 46 points, 32 comments]
• “Stop treating ‘AGI’ as the north-star goal of AI research,” has been accepted to #ICML2025! arxiv.org/abs/2502.03689 [bsky, 27 points, 2 comments]
• Stop treating `AGI' as the north-star goal of AI research [lobsters, 22 points, 5 comments]
• feels like you might enjoy “Stop treating 'AGI' as the north-star goal of AI research” c/o @struthious.bsky.social, @taliaringer.bsky.social, et al. arxiv.org/abs/2502.03689 [bsky, 17 points, 2 comments]
• I enjoyed this interview. It is refreshing to find critical views of AI! I look forward to reading the paper: arxiv.org/abs/2502.03689 [bsky, 14 points, 0 comments]
• Strong critiques of AGI in here. My question: is AGI is a genuine movitation for most AI progress, or more of a justification? @borhane.bsky.social @eryk.bsky.social @mmitchell.bsky.social arxiv.org/a [bsky, 11 points, 2 comments]
• arxiv.org/abs/2502.03689 [bsky, 6 points, 1 comments]
• arxiv.org/abs/2502.03689 [bsky, 5 points, 0 comments]
• We'll get to AGI *someday* but it's not happening anytime soon. Sam Altman is not going to get his pet Omnissiah. But it remains true that the gap is visibly narrowing and we've built very parallel-c [bsky, 5 points, 0 comments]
• Great article pushing back against AGI hype: "Rather than a single general north-star goal (or small set of goals), the AI community should articulate many worthwhile scientific engineering, and socie [bsky, 4 points, 0 comments]
• Ooh! A friend just led me onto a new paper. arxiv.org/abs/2502.03689 [bsky, 3 points, 0 comments]
• And this one by @borhane.bsky.social et al: arxiv.org/pdf/2502.036... [bsky, 2 points, 0 comments]
• Marketinggedreven verhalen zijn overal en spelen een grote rol. De huidige hype rondom AI is daar een perfect voorbeeld van. Hier is een wat gematigder artikel, zonder de gebruikelijke overdrijvingen: [bsky, 2 points, 0 comments]
• Directly addresses some of the most grating things about the AI boom from a research perspective: the lack of specific and focused goals, the lack of distinction between engineering and science, the i [bsky, 1 points, 1 comments]
• Stop treating `AGI' as the north-star goal of AI research https://arxiv.org/abs/2502.03689 (https://news.ycombinator.com/item?id=43876843) [bsky, 0 points, 0 comments]
• arxiv.org/pdf/2502.03689 [bsky, 0 points, 0 comments]
• Stop treating `AGI' as the north-star goal of AI research https://lobste.rs/s/kfadjc #ai [bsky, 0 points, 0 comments]
• https://bsky.app/profile/lobste.rs.web.brid.gy/post/3lobiadhg7lq2 [bsky, 0 points, 0 comments]

Related

• Why We Don't Have AGI Yet
• What Is Artificial General Intelligence?
• AI Must Embrace Specialization via Superhuman Adaptable Intelligence
• From AGI to ASI
• Unsocial Intelligence: an Investigation of the Assumptions of AGI Discourse