技术公司转而使用人工智能培训的合成数据,但专家对质量和准确性风险提出警告。 Tech firms turn to synthetic data for AI training, but experts warn of quality and accuracy risks.

由于人造数据短缺,技术公司越来越多地使用算法创造的合成数据来培训人工智能模型。 Tech companies are increasingly using synthetic data, created by algorithms, to train AI models due to a shortage of human-generated data. 合成数据虽然成本更低、速度更快, 却可以导致人工智能“焚化”和过于简单化的产出。 While it's cheaper and faster, synthetic data can lead to AI "hallucinations" and overly simplistic outputs. 专家建议采用全球标准跟踪和验证这些数据,以确保质量并保持AI的准确性和可信度。 Experts suggest global standards for tracking and validating this data are needed to ensure quality and maintain AI accuracy and trustworthiness.