The research of the Massachusetts Institute of Technology finds that AI has no stable values.

MIT Study Finds That AI Doesn't In Fact Have Stable Values

A few months ago, a study suggested that as artificial intelligence becomes more complex, it could develop a "value system," such as prioritizing its own well - being over that of humans. However, a recent paper published by the Massachusetts Institute of Technology (MIT) has poured cold water on this exaggerated view, concluding that AI actually doesn't have any coherent values. The co - authors of the MIT study said their findings suggest that "aligning" AI systems may be more challenging than commonly thought. They emphasized that AI as we know it today can "hallucinate" and imitate, making it difficult to predict in many ways. They said that no model is consistent in its preferences. Depending on the wording and framing of the prompts, they can exhibit very different views.

—— TechCrunch

via Wind Vane Reference Express - Telegram Channel