AI Might Now Be as Good as Humans at Detecting Emotion, Political Leaning, and Sarcasm
1 min read
Summary
A recent study has shown that large language models (LLMs) are as good as humans at analysing sentiment, political stance, emotional intensity and sarcasm detection in online messaging.
The research analysed 33 human subjects and 100 pieces of text, evaluating the reliability, consistency and quality of seven LLMs, including GPT-4, Gemini, Llama-3.1-70B and Mixtral 8 × 7B.
It found that LLMs were more consistent than humans at spotting political leanings and intensities of emotions, but struggled with detecting sarcasm.
The study suggests that LLMs could be utilised to dramatically cut the time and cost of analysing large amounts of online content, such as tweets.
However, it did highlight that more work needs to be done to ensure the outputs of LLMs remain consistent.
