Epistemic diversity across language models mitigates knowledge collapse
This addresses the risk of AI monoculture for AI developers and policymakers by showing that model diversity can help maintain performance, though it is incremental as it builds on prior single-model collapse research.
The study tackled the problem of knowledge collapse in AI by investigating whether diversity among language models can mitigate performance decay when models are trained on their own collective output, finding that increased epistemic diversity reduces collapse but only up to an optimal level, with performance decaying if models are too few or too many.
The growing use of artificial intelligence (AI) raises concerns of knowledge collapse, i.e., a reduction to the most dominant and central set of ideas. Prior work has demonstrated single-model collapse, defined as performance decay in an AI model trained on its own output. Inspired by ecology, we ask whether AI ecosystem diversity, that is, diversity among models, can mitigate such a collapse. We build on the single-model approach but focus on ecosystems of models trained on their collective output. To study the effect of diversity on model performance, we segment the training data across language models and evaluate the resulting ecosystems over ten, self-training iterations. We find that increased epistemic diversity mitigates collapse, but, interestingly, only up to an optimal level. Our results suggest that an ecosystem containing only a few diverse models fails to express the rich mixture of the full, true distribution, resulting in rapid performance decay. Yet distributing the data across too many models reduces each model's approximation capacity on the true distribution, leading to poor performance already in the first iteration step. In the context of AI monoculture, our results suggest the need to monitor diversity across AI systems and to develop policies that incentivize more domain- and community-specific models.