New Research Shows How to Make AI Judgments More Human-Like
Scientists found simple ways to ask AI models that make their judgments more stable and accurate. This could help AI give better advice on moral and social questions. The key is using the right prompts when asking the AI.
Researchers from arXiv published a study showing how to make AI models better at understanding human judgments. They discovered that by using simple, well-structured prompts, AI models can give more consistent and accurate answers to moral and social questions. The study tested these techniques on two datasets: a U.S.-representative set of 144 moral scenarios and 38 moral beliefs from the International Social Survey Programme's Family and Changing Gender Roles module covering 32 countries.
This matters because AI often struggles to match human judgment, especially when the wording of a question changes slightly. For example, an AI might give different answers to the same question if you rephrase it. The new techniques help AI provide more stable and reliable responses, making it more useful for real-world applications like advice-giving and decision-making.
If you're curious, you can try this out today. Open any AI chat tool like ChatGPT or Claude and ask a moral or social question in different ways. Notice how the answers vary. Then, try rephrasing your question using clear, straightforward language. You might see more consistent responses, just like the researchers did.