In a blog post, Microsoft researchers detailed the potential of using synthetic data to train artificial intelligence models. Use cases are currently limited to short text passages, but the researchers argued synthetic data could be a positive step in creating more training material for AI without compromising privacy.
Full story