Can Robots Write Treaties? Using Recurrent Neural Networks to Draft International Investment Agreements
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Negotiating international investment agreements is costly, complex, and prone to power asymmetries. Would it then not make sense to let computers do part of the work? In this contribution, we train a character-level recurrent neural network (RNN) to write international investment agreements. Benefitting from the formulaic nature of treaty language, the RNN generates texts of lawyer-like quality on the article-level, but fails to compose treaties in a legally sensible manner. By embedding RNNs in a user-controlled pipeline we overcome this problem. First, users can specify the treaty content categories ex ante on which the RNN is trained. Second, the pipeline allows a filtering of output ex post by identifying output that corresponds most closely to a user-selected treaty design benchmark. The result is an improved system that produces meaningful texts with legally sensible composition. We test the pipeline by comparing predicted treaties to actually concluded ones and by verifying that our filter captures latent policy preferences by predicting the outcome of current investment treaty negotiations between China and the United States.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it