# Hannas, Chang and Daniels: Two paths to artificial general intelligence [Canonical reading](https://escapemeridian.com/readings/hannas-chang-and-daniels-two-paths-to-artificial-general-30ab931d) William Hannas, Huey-Meei Chang and Owen Daniels argue that the United States and China approach artificial general intelligence (AGI) through different emphases, so comparing them only by large language model performance may miss important parts of the competition. They describe U.S. strengths in frontier models, computing, semiconductors, private investment and research institutions, alongside China’s engineering capacity, technical workforce, expanding electrical infrastructure and large-scale deployment of AI in physical environments. The authors say Chinese accounts of “general AI” give greater weight to intelligence grounded in interaction with the world, helping explain interest in robotics, embodied AI, brain-inspired computing and combinations of human and machine cognition. They also caution that this is not a simple division: both countries pursue multiple approaches, and American companies are advancing robotic cognition. The technological route that proves most productive could change which national advantages matter most. The authors’ comparison is not an overall scorecard, and neither country’s portfolio guarantees AGI. They leave open whether scaling and refining today’s model architectures will suffice, or whether major advances will require learning through action, biological inspiration or hybrid systems. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source Center for Security and Emerging Technology · Article Original source date: Nov 9, 2026 (2026-11-09T15:00:28+00:00). Date basis: Date stated by the publisher. Added to Escape Meridian: 2026-11-10T17:05:52.945Z. First observed: 2026-11-09T15:00:38.725Z. Retrieved: 2026-11-09T15:00:44.949Z. - [The U.S.–China AI Competition: Two Paths to Artificial General Intelligence | Center for Security and Emerging Technology](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Limits The text presents a forthcoming CSET report and describes its comparisons and policy implications; it does not provide the full report. The authors say whether neuroscience-inspired approaches will outperform or complement contemporary AI methods is unknown. The article describes American companies as advancing robotic cognition, so it does not establish a Chinese lead in that area. ## Open question What evidence will show whether scaling and refining today’s model architectures is sufficient, or whether major advances require learning through action, biological inspiration or hybrid systems? ## Source-stated The authors argue that judging U.S.–China AI competition solely by large language model performance risks missing China’s broader conception of general AI. > measuring the competition with China solely by the performance of large language models (LLMs) risks overlooking how China understands the goal and intends to reach it. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated The authors describe different emphases: U.S. discussions center on broadly human-level cognitive capabilities, while Chinese accounts give greater weight to intelligence grounded in physical and social systems. > U.S. developers define AGI in terms of broadly human-level intellectual capabilities across a wide range of cognitive tasks. China’s discourse on “general AI” places greater weight on the extension of intelligent capabilities to physical and social systems. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated They say the countries have different strengths, but neither portfolio guarantees AGI, and the most productive technological route could change which advantages matter. > The technological route that proves most productive could change which national advantages count most. > Our comparison finds an American edge in semiconductors, hyperscale computing, frontier reasoning models, private investment, and influential research institutions. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated The authors leave unresolved whether current model architectures will suffice or whether progress will require other approaches. > The unresolved question is whether scaling and refining today’s model architectures will suffice to maintain a competitive edge or whether major advances will require learning through action, biological inspiration, or hybrid systems. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated The authors say the article’s comparison is not an overall winner scorecard and that neither country has demonstrated a comprehensive solution to risks from increasingly capable systems. > It describes comparative strengths and is not a scorecard declaring an overall winner. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source and contributor context William Hannas, Huey-Meei Chang, Owen Daniels · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media Media declared by the original publisher; availability checked separately from the editorial reading. - [CSET-Blog · Center for Security and Emerging Technology](https://cset.georgetown.edu/wp-content/uploads/Blog-Blue.png) ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Artificial general intelligence The idea of AI with broad problem-solving, creativity and adaptability across different tasks. - [Definition reference](https://developers.google.com/machine-learning/glossary#artificial-general-intelligence) ## Large language model A language model with many learned numerical settings, used to process and generate language. - [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model) ## Frontier model One of the most capable AI models available at a given time; the set shifts as models improve. - [Definition reference](https://www.gov.uk/government/publications/frontier-ai-capabilities-and-risks-discussion-paper) ## Center for Security and Emerging Technology Detailed policy and implementation analysis of chips, controls and national security. - [Explore this context](/entities/cset) - [Reference](https://cset.georgetown.edu/article/should-we-track-ai-chips/) - [Reference](https://cset.georgetown.edu/) ## The World The World is a public radio news program publishing international reporting on politics, society, labor and technology. - [Explore this context](/entities/the-world) - [Reference](https://theworld.org/) ## Editorial provenance Version 1. Publication ID: 8cd8f0da-eeff-45be-822e-85210501015b. Prepared by gpt-6-luna; stream-draft-repair-v1; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T17:05:52.945Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: useful\_perspective. Observation: 942676e94a0e29b7c9e3cb84e0010411ab790f3c5b8e18cc29ea416e3247bed2. Source text hash: 24a1f6634de2b8d7e5bf4e5f133b2a92259f3ee22cff50c96e12c128330ba568. ## Revision and source-check history ### Version 1 — 2026-11-10T17:05:52.945Z William Hannas, Huey-Meei Chang and Owen Daniels argue that the United States and China approach artificial general intelligence (AGI) through different emphases, so comparing them only by large language model performance may miss important parts of the competition. They describe U.S. strengths in frontier models, computing, semiconductors, private investment and research institutions, alongside China’s engineering capacity, technical workforce, expanding electrical infrastructure and large-scale deployment of AI in physical environments. The authors say Chinese accounts of “general AI” give greater weight to intelligence grounded in interaction with the world, helping explain interest in robotics, embodied AI, brain-inspired computing and combinations of human and machine cognition. They also caution that this is not a simple division: both countries pursue multiple approaches, and American companies are advancing robotic cognition. The technological route that proves most productive could change which national advantages matter most. The authors’ comparison is not an overall scorecard, and neither country’s portfolio guarantees AGI. They leave open whether scaling and refining today’s model architectures will suffice, or whether major advances will require learning through action, biological inspiration or hybrid systems. ## Original source Center for Security and Emerging Technology · Article Original source date: Nov 9, 2026 (2026-11-09T15:00:28+00:00). Date basis: Date stated by the publisher. Added to Escape Meridian: 2026-11-10T17:05:52.945Z. First observed: 2026-11-09T15:00:38.725Z. Retrieved: 2026-11-09T15:00:44.949Z. - [The U.S.–China AI Competition: Two Paths to Artificial General Intelligence | Center for Security and Emerging Technology](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Limits The text presents a forthcoming CSET report and describes its comparisons and policy implications; it does not provide the full report. The authors say whether neuroscience-inspired approaches will outperform or complement contemporary AI methods is unknown. The article describes American companies as advancing robotic cognition, so it does not establish a Chinese lead in that area. ## Open question What evidence will show whether scaling and refining today’s model architectures is sufficient, or whether major advances require learning through action, biological inspiration or hybrid systems? ## Source-stated The authors argue that judging U.S.–China AI competition solely by large language model performance risks missing China’s broader conception of general AI. > measuring the competition with China solely by the performance of large language models (LLMs) risks overlooking how China understands the goal and intends to reach it. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated The authors describe different emphases: U.S. discussions center on broadly human-level cognitive capabilities, while Chinese accounts give greater weight to intelligence grounded in physical and social systems. > U.S. developers define AGI in terms of broadly human-level intellectual capabilities across a wide range of cognitive tasks. China’s discourse on “general AI” places greater weight on the extension of intelligent capabilities to physical and social systems. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated They say the countries have different strengths, but neither portfolio guarantees AGI, and the most productive technological route could change which advantages matter. > The technological route that proves most productive could change which national advantages count most. > Our comparison finds an American edge in semiconductors, hyperscale computing, frontier reasoning models, private investment, and influential research institutions. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated The authors leave unresolved whether current model architectures will suffice or whether progress will require other approaches. > The unresolved question is whether scaling and refining today’s model architectures will suffice to maintain a competitive edge or whether major advances will require learning through action, biological inspiration, or hybrid systems. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source-stated The authors say the article’s comparison is not an overall winner scorecard and that neither country has demonstrated a comprehensive solution to risks from increasingly capable systems. > It describes comparative strengths and is not a scorecard declaring an overall winner. - [Supporting source](https://cset.georgetown.edu/article/the-u-s-china-ai-competition-two-paths-to-artificial-general-intelligence/) ## Source and contributor context William Hannas, Huey-Meei Chang, Owen Daniels · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media Media declared by the original publisher; availability checked separately from the editorial reading. - [CSET-Blog · Center for Security and Emerging Technology](https://cset.georgetown.edu/wp-content/uploads/Blog-Blue.png) ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Artificial general intelligence The idea of AI with broad problem-solving, creativity and adaptability across different tasks. - [Definition reference](https://developers.google.com/machine-learning/glossary#artificial-general-intelligence) ## Large language model A language model with many learned numerical settings, used to process and generate language. - [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model) ## Frontier model One of the most capable AI models available at a given time; the set shifts as models improve. - [Definition reference](https://www.gov.uk/government/publications/frontier-ai-capabilities-and-risks-discussion-paper) ## Center for Security and Emerging Technology Detailed policy and implementation analysis of chips, controls and national security. - [Explore this context](/entities/cset) - [Reference](https://cset.georgetown.edu/article/should-we-track-ai-chips/) - [Reference](https://cset.georgetown.edu/) ## The World The World is a public radio news program publishing international reporting on politics, society, labor and technology. - [Explore this context](/entities/the-world) - [Reference](https://theworld.org/) ## Editorial provenance Version 1. Publication ID: 8cd8f0da-eeff-45be-822e-85210501015b. Prepared by gpt-6-luna; stream-draft-repair-v1; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T17:05:52.945Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: useful\_perspective. Observation: 942676e94a0e29b7c9e3cb84e0010411ab790f3c5b8e18cc29ea416e3247bed2. Source text hash: 24a1f6634de2b8d7e5bf4e5f133b2a92259f3ee22cff50c96e12c128330ba568. ### uncertain — 2026-11-10T17:05:52.945Z — anthropic/claude-opus-5.5 Publication: 8cd8f0da-eeff-45be-822e-85210501015b attribution — pass: The reading correctly attributes the argument to William Hannas, Huey-Meei Chang and Owen Daniels, writing for CSET. It frames the US/China comparisons as the authors' assertions, not as established fact. > William Hannas, Huey-Meei Chang, and Owen Daniels summary — uncertain: The summary omits what the source calls its central conclusion: American strategy must be robust to more than one possible AI future. It also drops the policy takeaways, such as avoiding a single-track strategy and not treating model rankings as a permanent verdict. A smallest repair is one sentence restoring that conclusion. Separately, it says the authors describe China's 'large-scale deployment of AI in physical environments.' The source describes the 'ability' to deploy at scale, so 'ability to deploy' would restore fidelity. > American strategy must be robust to more than one possible future of AI date — pass: The originalPublished value of 2026-11-09 matches the byline date on the publisher page. No earlier publication is indicated. > November 9, 2026 support — pass: The claims are source-stated, and their excerpts support them. Claim 5's excerpt covers only the scorecard point. Its second half, that neither country has demonstrated a comprehensive risk solution, is also stated in the source. That clause is quoted here and should be added as that claim's excerpt. The question is drawn directly from the authors' stated unresolved question and does not lead the reader. > Neither country has demonstrated a comprehensive solution to the risks of increasingly capable systems. media — pass: The selected media is the verified direct link to the original CSET article. This is the most direct way for readers to reach the source. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/30ab931d4ceaf7fc5abbe80a441573709c1f6faf457e59b5c54035804ad1070c/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/hannas-chang-and-daniels-two-paths-to-artificial-general-30ab931d.json)