A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates a significant advancement in making AI systems more interpretable. Researchers developed a method that allows large language models to explain the reasoning behind their decisions in a step-by-step manner, similar to a human showing their work. This technique, applied to models performing complex question-answering and code-generation tasks, generates internal “scratchpads” that reveal the model’s logical progression. The research aims to address the “black box” problem in AI, where it’s difficult to understand why a model produces a specific output. While promising for improving trust and debugging in critical applications, the authors note that the explanations themselves require validation and the approach adds computational overhead. For the full details, read the complete article at https://technologyreview.com/2024/05/15/1090000/mit-ai-explainability-breakthrough.