Mittelsteadt adds that Trump could punish companies in various ways. For instance, his government canceled a significant federal contract with Amazon Web Services, influenced by his perception of the Washington Post's Jeff Bezos.
A 2023 study from the University of Washington and others found diverse political leanings in large language models, impacting hate speech and misinformation detection performance due to political bias.
Yejin Bang notes that while most AI models lean liberal and US-centric, they can exhibit both liberal and conservative biases based on the topic, highlighting inconsistency in AI responses.
Ashique KhudaBukhsh emphasizes that as AI systems proliferate, the subtle introduction of biases through model training can have increasingly negative consequences for public discourse.
Collection
[
|
...
]