Brynjolfsson versus Gordon on Artificial Intelligence and Productivity Growth: What the Science Says
Bibliographic record
Abstract
Over the past decade, Erik Brynjolfsson and Robert Gordon have debated each other over the question of Artificial Intelligence (AI) and Productivity Growth, with the former predicting robust, sustained growth and the latter casting doubt over the ability of AI to replicate the growth rates of the 20th century. Until now, the debate has been largely anecdotal, with little scientific content. Brynjolfsson has focused on a number of promising forms of AI, while Gordon has raised a number of headwinds. This paper attempts to further the debate by invoking the principles of basic material processes. In other words, what does science have to say regarding the role of AI in productivity growth? To this end, it examines AI from the point of view of process engineering as well as various econophysics-based models of economic growth. The evidence presented provides support for Robert Gordon’s position, but for entirely different reasons. Specifically, it is shown that because information is not physically productive, more and better information (obtained from machine learning and other algorithms) cannot and will not increase productivity and, as such, cannot increase growth. There are, however, exceptions. For example, more and better information can contribute to increasing second-law efficiency, triggering one-shot increases in output. In this regard, AI is seen as being analogous to the information and communications technology revolution in the 1980s and 1990s. Our results predict the AI equivalent of the information paradox, namely that in the future, you will see AI everywhere but in the growth statistics.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.025 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.003 | 0.005 |
| Science and technology studies | 0.003 | 0.014 |
| Scholarly communication | 0.008 | 0.023 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.010 | 0.010 |
| Insufficient payload (model declined to judge) | 0.006 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".