A team from the University of Geneva (UNIGE) and the University of Bern (UniBE) conducted a study into AI's understanding of human emotions and claimed they outperform humans;
It's a very arresting idea, but I need an expert in LLMs and the way they work to comment on the finding while keeping the following in mind;
• The methods are key in the cause for skepticism-- the way they used multiple choice questions to simulate the 'conflict' is slightly amusing, and conjures a pretty funny image of somebody interrupting an emotionally charged situation and attempting to handle it in a way that's akin to a taking multiple choice test
• Does the core claim ultimately stand up once you consider the methods and results? Does the claim of the headline stand up to scrutiny?
posted6/5/2025
deadline6/10/2025
processing
published9/23/2025
Recently published by Live Science
I'm writing news coverage of the new ChatGPT agent from OpenAI and need an AI expert to comment. It's vital that the source NOT be an AI vendor or anyone that currently works for a company selling any kind of AI product, nor anyone associated directly with OpenAI in any way. My questions for the expert are:
Is this legit? Can the new agent deliver on the promises OpenAI is making for it?
What are the risks of the system?
A team from the University of Geneva (UNIGE) and the University of Bern (UniBE) conducted a study into AI's understanding of human emotions and claimed they outperform humans;
It's a very arresting idea, but I need an expert in LLMs and the way they work to comment on the finding while keeping the following in mind;
• The methods are key in the cause for skepticism-- the way they used multiple choice questions to simulate the 'conflict' is slightly amusing, and conjures a pretty funny image of somebody interrupting an emotionally charged situation and attempting to handle it in a way that's akin to a taking multiple choice test
• Does the core claim ultimately stand up once you consider the methods and results? Does the claim of the headline stand up to scrutiny?