Stop treating ‘AGI’ as the north-star goal of AI research

Reference. Stop treating ‘AGI’ as the north-star goal of AI research

Borhane Blili-Hamelin, Christopher Graziul, Leif Hancox-Li, Hananel Hazan, El-Mahdi El-Mhamdi, Avijit Ghosh, Katherine Heller, Jacob Metcalf, Fabricio Murai, Eryk Salvaggio, Andrew Smart, Todd Snider, Mariame Tighanimine, Talia Ringer, Margaret Mitchell, Shiri Dori-Hacohen · arXiv · 2025 · machine-learning · arXiv

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.

Cite

Cite as @blilihamelin-2025-stop (helia, typst) · \cite{blilihamelin-2025-stop} (LaTeX)

BibTeX

bibtex · 8 lines

@misc{blilihamelin-2025-stop,
  author = {Borhane Blili-Hamelin and Christopher Graziul and Leif Hancox-Li and Hananel Hazan and El-Mahdi El-Mhamdi and Avijit Ghosh and Katherine Heller and Jacob Metcalf and Fabricio Murai and Eryk Salvaggio and Andrew Smart and Todd Snider and Mariame Tighanimine and Talia Ringer and Margaret Mitchell and Shiri Dori-Hacohen},
  title = {Stop treating ‘AGI’ as the north-star goal of AI research},
  year = {2025},
  month = {2},
  eprint = {2502.03689},
  archiveprefix = {arXiv}
}

hayagriva YAML (typst)

yaml · 23 lines

blilihamelin-2025-stop:
  type: misc
  title: Stop treating ‘AGI’ as the north-star goal of AI research
  author:
  - Blili-Hamelin, Borhane
  - Graziul, Christopher
  - Hancox-Li, Leif
  - Hazan, Hananel
  - El-Mhamdi, El-Mahdi
  - Ghosh, Avijit
  - Heller, Katherine
  - Metcalf, Jacob
  - Murai, Fabricio
  - Salvaggio, Eryk
  - Smart, Andrew
  - Snider, Todd
  - Tighanimine, Mariame
  - Ringer, Talia
  - Mitchell, Margaret
  - Dori-Hacohen, Shiri
  date: 2025-02
  serial-number:
    arxiv: '2502.03689'

Download: blilihamelin-2025-stop.bib · blilihamelin-2025-stop.yml

Cites 222 works (0 here)

External (222)

A metatheory of classical and modern connectionism (2025)
How should we test AI for human-level intelligence? OpenAI’s o3 electrifies quest (2025)
Use of Exclusive Data for Corporate Research on Machine Learning and Artificial Intelligence: Implications for Innovation and Competition Policy (2025)
Opinion | The Government Knows A.G.I. Is Coming (2025)
Agi is coming faster than we think: We must get ready now (2025)
Trump to announce private sector ai infrastructure investment (2025)
Anthropic's recommendations to OSTP for the U.S. AI Action Plan (2025)
Draft Report of the Joint California Policy Working Group on AI Frontier Models (2025)
About (Google DeepMind) (2025)
AAAI 2025 presidential panel on the future of AI research (2025)
Planning for AGI and beyond (2025)
Security on the path to AGI (2025)
Our structure (OpenAI) (2025)
AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How to Tell the Difference (2024)
ARC Prize 2024: Technical Report (2024)
Reclaiming AI as a Theoretical Tool for Cognitive Science (2024)
SPEECH RECOGNITION FOR ANALYSIS OF POLICE RADIO COMMUNICATION (2024)
Advancing cybersecurity: a comprehensive review of AI-driven detection techniques (2024)
Light bulbs have energy ratings — so why can’t AI chatbots? (2024)
The reanimation of pseudoscience in machine learning and its ethical repercussions (2024)
Benchmarks as Microscopes: A Call for Model Metrology (2024)
Position: Measure Dataset Diversity, Don't Just Claim It (2024)
ChatGPT is bullshit (2024)
Testing learning hypotheses using neural networks by manipulating learning data (2024)
Race and Privacy in Broadcast Police Communications (2024)
STAR: SocioTechnical Approach to Red Teaming Language Models (2024)
Quantifying Misalignment Between Agents: Towards a Sociotechnical Understanding of Alignment (2024)
Science in the Age of AI (2024)
Reducing Biases towards Minoritized Populations in Medical Curricular Content via Artificial Intelligence for Fairer Health Outcomes (2024)
Real Risks of Fake Data: Synthetic Data, Diversity-Washing and Consent Circumvention (2024)
Position: Why We Must Rethink Empirical Research in Machine Learning (2024)
A Careful Examination of Large Language Model Performance on Grade School Arithmetic (2024)
Benchmarking Benchmark Leakage in Large Language Models (2024)
Field-building and the epistemic culture of AI safety (2024)
Participation versus scale: Tensions in the practical demands on participatory AI (2024)
The TESCREAL bundle: Eugenics and the promise of utopia through artificial general intelligence (2024)
Staggered deployment of gunshot detection technology in Chicago, IL: a matched quasi-experiment of gun violence outcomes (2024)
STELA: a community-centred approach to norm elicitation for AI alignment (2024)
Epistemic Power in AI Ethics Labor: Legitimizing Located Complaints (2024)
A Roadmap to Pluralistic Alignment (2024)
Position: Stop Making Unscientific AGI Performance Claims (2024)
Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs (2024)
Data work and practices in healthcare: A scoping review (2024)
Unsocial Intelligence: An Investigation of the Assumptions of AGI Discourse (2024)
How the AI Boom Went Bust (2024)
Against Predictive Optimization: On the Legitimacy of Decision-making Algorithms That Optimize Predictive Accuracy (2024)
When Might AI Outsmart Us? It Depends Who You Ask (2024)
Mirror, Mirror on the Wall: Algo-rithmic Assessments, Transparency, and Self-Fulﬁlling Prophecies (2024)
OpenAI o3 Breakthrough High Score on ARC-AGI-Pub (2024)
Debates on the nature of artificial general intelligence (2024)
GPT-4 technical report (2024)
Position: A call for embodied AI (2024)
Using large language models to promote health equity (2023)
Position: Levels of AGI for Operationalizing Progress on the Path to AGI (2023)
Language Models Represent Space and Time (2023)
The Participatory Turn in AI Design: Theoretical Foundations and the Current State of Practice (2023)
Reimagining Democracy for AI (2023)
Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties (2023)
AI safety on whose terms? (2023)
The uncontroversial ‘thingness’ of AI (2023)
Police Use of Facial Recognition Technology and Racial Bias – An Assessment of Criticisms of Its Current Use (2023)
Algorithmic Pluralism: A Structural Approach To Equal Opportunity (2023)
Surprising combinations of research contents and contexts are related to impact and emerge with scientific outsiders from distant disciplines (2023)
Evaluation for Change (2023)
The Coming Wave (2023)
Characterization of Stigmatizing Language in Medical Records (2023)
US Justice Department Urged to Investigate Gunshot Detector Purchases (2023)
Artificial General Intelligence Is Already Here (2023)
Queering intelligence: A theory of intelligence as performance and a critique of individual and artificial intelligence (2023)
Sparks of artificial general intelligence: Early experiments with GPT-4 (2023)
Getting ready for artificial general intelligence with examples (2023)
Artificial general intelligence (AGI) is one prompt away (2023)
Does big data serve policy? Not without context. An experiment with in silico social science (2022)
Picking on the Same Person: Does Algorithmic Monoculture lead to Outcome Homogenization? (2022)
Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction (2022)
On the Impossible Safety of Large AI Models (2022)
Dislocated accountabilities in the “AI supply chain”: Modularity and developers’ notions of responsibility (2022)
Power to the People? Opportunities and Challenges for Participatory AI (2022)
Paradigm, specialty, pragmatism: Kuhn's legacy to methodological pluralism (2022)
Making Intelligence: Ethical Values in IQ and ML Benchmarks (2022)
Current and Near-Term AI as a Potential Existential Risk Factor (2022)
The Fallacy of AI Functionality (2022)
Large language models are few-shot clinical information extractors (2022)
Evaluation Gaps in Machine Learning Practice (2022)
Should attention be all we need? The epistemic and ethical implications of unification in machine learning (2022)
A Systematic Review and Thematic Analysis of Community-Collaborative Approaches to Computing Research (2022)
To ArXiv or not to ArXiv: A Study Quantifying Pros and Cons of Posting Preprints Online (2022)
Minding the gap(s): public perceptions of AI and socio-technical imaginaries (2022)
Whose AI Dream? In search of the aspiration in data annotation (2022)
The Worst of Both Worlds: A Comparative Analysis of Errors in Learning from Data in Psychology and Machine Learning (2022)
Compute Trends Across Three Eras of Machine Learning (2022)
Emerging Trends: SOTA-Chasing (2022)
Democratising Measurement: or Why Thick Concepts Call for Coproduction (2022)
Flat teams drive scientific innovation (2022)
A Proposal For The Dartmouth Summer Research Project On Artificial Intelligence (2022)
The transformative potential of artificial intelligence (2022)
High-Resolution Image Synthesis with Latent Diffusion Models (2021)
AI and the Everything in the Whole Wide World Benchmark (2021)
General intelligence disentangled via a generality metric for natural and artificial intelligence (2021)
Towards artificial general intelligence via a multimodal foundation model (2021)
Fairness via AI: Bias Reduction in Medical Information (2021)
Do Datasets Have Politics? Disciplinary Values in Computer Vision Dataset Development (2021)
AI in Finance: Challenges, Techniques, and Opportunities (2021)
The Benchmark Lottery (2021)
On the genealogy of machine learning datasets: A critical history of ImageNet (2021)
The Values Encoded in Machine Learning Research (2021)
A Silicon Valley love triangle: Hiring algorithms, pseudo-science, and the quest for auditability (2021)
What’s wrong with epistemic trespassing? (2021)
Getting Ourselves Together: Data-centered participatory design research & epistemic burden (2021)
Impact of ShotSpotter Technology on Firearm Homicides and Arrests Among Large Metropolitan Counties: a Longitudinal Analysis (2021)
On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜 (2021)
On the relationship between interdisciplinarity and impact: Distinct effects on academic and broader impact (2021)
Algorithmic monoculture and social welfare (2021)
Are We Learning Yet? A Meta Review of Evaluation Failures Across Machine Learning (2021)
A Relational Theory of Data Governance (2021)
The (Im)possibility of Fairness: Different Value Systems Require Different Mechanisms For Fair Decision Making (2021)
Design Justice: Community-Led Practices to Build the Worlds We Need (2020)
Underspecification Presents Challenges for Credibility in Modern Machine Learning (2020)
Towards decolonising computational sciences (2020)
Collaborative Learning in the Jungle (Decentralized, Byzantine, Heterogeneous, Asynchronous and Nonconvex Learning) (2020)
Interdisciplinarity and Team Innovation: The Role of Team Experiential and Relational Resources (2020)
A No-Free-Lunch Theorem for MultiTask Learning (2020)
Breaking Down the Silos: Innovations for Multidisciplinary Programs (2020)
The hardware lottery (2020)
Language (Technology) is Power: A Critical Survey of “Bias” in NLP (2020)
How many preprints have actually been printed and why: a case study of computer science preprints on arXiv (2020)
Credibility of preprints: an interdisciplinary survey of researchers (2020)
The Problem with Intelligence: Its Value-Laden History and the Future of AI (2020)
Artificial Intelligence, Values, and Alignment (2020)
Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing (2020)
Value-laden disciplinary shifts in machine learning (2020)
Law enforcement use of facial recognition: bias, disparate impacts on people of color, and the need for federal legislation (2020)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding (2019)
Measurement and Fairness (2019)
Roles for computing in social change (2019)
Dissecting racial bias in an algorithm used to manage the health of populations (2019)
Epistemic Trespassing (2019)
Model Pluralism (2019)
Artificial Intelligence and Journalism (2019)
The New Statistics for Better Science: Ask How Much, How Uncertain, and What Else is Known (2019)
Fairness and Abstraction in Sociotechnical Systems (2019)
Beyond Open vs. Closed: Balancing Individual Privacy and Public Accountability in Data Sharing (2019)
Spearman’s g Found in 31 Non-Western Nations: Strong Evidence That g Is a Universal Phenomenon (2019)
How can we ensure visibility and diversity in research contributions? How the Contributor Role Taxonomy (CRediT) is helping the shift from authorship to contributorship (2019)
On the Measure of Intelligence (2019)
Paradigms of Artificial General Intelligence and Their Associated Risks (2019)
AIs Versus Four-Year-Olds (2019)
Data Science as Political Action: Grounding Data Science in a Politics of Justice (2018)
Values and Ethics in Human-Computer Interaction (2018)
Gender Shades: In-tersectional Accuracy Disparities in Commercial Gender Classiﬁcation (2018)
Algorithmic Accountability and Public Reason (2017)
What to do when machines do everything: How to get ahead in a world of ai, algorithms, bots, and big data (2017)
Privacy is an essentially contested concept: a multi-dimensional analytic for mapping privacy (2016)
Trustworthiness and authority of scholarly information in a digital age: Results of an international questionnaire (2016)
Beyond Zero and One: Machines, Psychedelics, and Consciousness (2016)
The Superiority of Economists (2015)
Measuring the effectiveness of scientiﬁc gatekeeping (2015)
Google apologises for Photos app’s racist blun-der (2015)
Superintelligence: Paths, Dangers, Strategies (2014)
Universal psychometrics: Measuring cognitive abilities in the machine kingdom (2014)
Machine Learning: The High Interest Credit Card of Technical Debt (2014)
Artiﬁcial general intelligence: concept, state of the art, and future prospects (2014)
Disciplinary dilemma: working across research silos is harder than it looks (2014)
Ethical Community-Engaged Research: A Literature Review (2013)
Diversity and the Division of Cognitive Labor (2013)
How Diversity Can Drive Innovation (2013)
Mapping the Landscape of Human-Level Artificial General Intelligence (2012)
The Architecture of Human-Like General Intelligence (2012)
A Reconsideration of Deweyan Democracy (Reprint from 1989) (2011)
The singularity: A philosophical analysis (2010)
Wozniak: Could a computer make a cup of coffee? (2010)
Path Dependence in the Production of Scientific Knowledge (2009)
Culture in global knowledge societies: knowledge cultures and epistemic cultures (2007)
Universal Intelligence: A Deﬁ-nition of Machine Intelligence (2007)
The Epistemology of Democracy (2006)
Toward a Critical Technical Practice: Lessons Learned in Trying to Reform AI (2006)
Groups of diverse problem solvers can outperform groups of high-ability problem solvers (2004)
Psychology's Status as a Scientific Discipline: Its Empirical Placement within an Implicit Hierarchy of the Sciences (2004)
Structural Holes and Good Ideas (2004)
Evaluating transdisciplinary science (2003)
Situated Knowledge and the Interplay of Value Judgments and Evidence in Scientific Inquiry (2002)
Addressing Social Determinants of Health: Community-Based Partici-patory Approaches to Research and Practice (2002)
Methodological Pluralism (2000)
Epistemic cultures : how the sciences make knowledge (1999)
BRINGING THE PEOPLE BACK IN (1997)
Nanotechnology and International Security (1997)
Path Dependence, Lock-In, and History (1995)
Some Philosophical Problems from the Standpoint of Artificial Intelligence (1981)
The Mismeasure of Man (1981)
Minds, brains, and programs (1980)
Mind over Machine (1978)
Computer power and human reason: from judgment to calculation (1976)
The search for generality (1965)
Natural general intelligence: how understanding the brain can help us build AI
AI that can match humans at any task will be here in ﬁve to 10 years
Most Researchers Do Not Believe AGI Is Imminent
Elon Musk reveals xAI efforts, predicts
Aligning artiﬁcial intelligence with climate change mitigation
Reasons for Action: Justiﬁcation, Motivation, Explanation
Machines of loving grace
Meta’s AI chief Yann LeCun
Human-Level Artiﬁcial Intelligence? Be Serious!
Disorder-ing Datasets: Sociotechnical Misalignments in AI-Mediated Behavioral Health
The myth of agi
The Dangers of Imposing Global North Approaches to AI Governance on the Global South | TechPolicy
”So, is this AGI?...”
On the diminishing returns to scaling
The Use of Artiﬁcial Intelligence in International Human Rights Law
Elon musk says ai will take your job, and ’no one is going to need to work’
Power and AI: Nature and Justiﬁcation
Innocent until Predicted Guilty: How Pre-mature Predictive Policing Can Lead to a Self-Fulﬁlling Prophecy of Juvenile Delinquency Note
Governing AI for Humanity. Final Report, United Nations, New York, NY, September
The democracy summit 2023
Watching the Generative AI Hype Bubble Deﬂate, August
J3016 202104: Taxonomy and Deﬁni-tions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles
microsoft-and-openai-have-a-ﬁnancial-deﬁnition-of-agi-report/
The ‘AI divide’ between the Global North and Global South
Dear elon musk, here are ﬁve things you might want to consider about agi
DeepSeek and the Race to AGI: How Global AI Competition Puts Ethical Accountability at Risk | TechPolicy.Press
Racist Technology in Action: Image recognition is still not capable of differentiating gorillas from Black people — racismandtechnol-ogy
Google’s Photo App Still Can’t Find Gorillas
Unrepro-ducible research is reproducible