Did AI get more negative recently?

TL;DR

This study develops a model to classify NLP and ML papers as positive or negative in stance, revealing that recent years have seen an increase in negative papers which are also more influential, contrasting with the overall trend of increasing positivity.

Contribution

The paper introduces a SciBERT-based model for automatic stance classification of AI research papers and analyzes long-term trends in paper sentiment over 35 years.

Findings

Papers have become more positive over time.

Negative papers have also become more negative recently.

Negative papers tend to receive more citations.

Abstract

In this paper, we classify scientific articles in the domain of natural language processing (NLP) and machine learning (ML), as core subfields of artificial intelligence (AI), into whether (i) they extend the current state-of-the-art by the introduction of novel techniques which beat existing models or whether (ii) they mainly criticize the existing state-of-the-art, i.e. that it is deficient with respect to some property (e.g. wrong evaluation, wrong datasets, misleading task specification). We refer to contributions under (i) as having a 'positive stance' and contributions under (ii) as having a 'negative stance' (to related work). We annotate over 1.5 k papers from NLP and ML to train a SciBERT-based model to automatically predict the stance of a paper based on its title and abstract. We then analyse large-scale trends on over 41 k papers from the last approximately 35 years in NLP and ML, finding that papers have become substantially more positive over time, but negative papers also got more negative and we observe considerably more negative papers in recent years. Negative papers are also more influential in terms of citations they receive.