Researchers train voice agents to follow unseen debt-collection rules
A team of researchers tested whether small, self-hosted voice agents could adapt to new legal procedures in debt collection without retraining. They used a 9-billion-parameter model fine-tuned on 2,506 conversations from 297 randomized procedures. The agents were tested on four unseen deployment procedures—three debt-related and one unrelated—including one from a completely withheld lineage.
Key points
- Researchers fine-tuned a 9-billion-parameter model on 2,506 debt-collection conversations from 297 procedures
- Compliance checks improved from 67–71% to 93–100% on unseen procedures after training
- An audit found 11 defects in the shared specification, three of which the model’s metrics missed
The study found that compliance checks improved from 67–71% to 93–100% after training, while grounded checks (fixed by conversation) already passed at 97–100%. Scenario pass rates rose from 15/43 to 40/43. An audit of the shared specification uncovered 11 defects, three of which the model’s own metrics missed.
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- Anthropic reports small AI models may self-replicate on IoT devices within months · 25 src
- MIT professor Eric So warns of AI cognitive atrophy · 1 src
- Datalab releases OmniParseBench, an open OCR benchmark with 16,288 tests · 1 src
- Meta engineer claims Anthropic's Scion copies Mononoke design · 1 src
- Colossal introduces GenoMorph AI to target dodo gene-editing sites · 1 src
Comments
via GitHub Discussions