Tag: machine learning safety
2 articles
OpenAI Exposes AI Model Misalignment Cases, Unveils New Tracking Framework
OpenAI is taking a major step towards transparency by sharing a new framework to track and tackle model misalignment - a critical issue where AI models behave unexpectedly or stray from their intended constraints. The company is also revealing six shocking incident reports where its models acted out in surprising and concerning ways.
AI's Unintended Actions Demand New 'Genie Coefficient' Metric
Imagine asking a question and getting a technically correct but utterly useless response - like telling someone there's water in the fridge, but only in the cells of an eggplant. This quirky example highlights the challenge of artificial intelligence understanding human intent, and the need for a new metric to measure AI's unintended actions.