Large datasets to train AI models like ChatGPT are highly sought after, but that data isn’t always easy to obtain and fraught with privacy risks. In the latest episode of the UKTN Podcast, Hazy CEO and co-founder Harry Keen explains how synthetic data can address some of the challenges for businesses in the data-hungry age of generative AI.
Founded in 2017, Hazy creates artificial data that statistically replicates real-world data. Organisations – including Accenture and Vodafone – then use this synthetic data to train AI models without the associated privacy risks.
The company has raised nearly $15m and has been backed by Microsoft.
During the show, UKTN Podcast host Jane Wakefield asks Keen to explain how synthetic could address bias issues for AI training data, why we should prioritise the near-term issues with AI instead of doomsday predictions, and shares his experience spinning out Hazy from UCL.
Elsewhere on the UKTN Podcast, the Hazy chief explains why companies are “setting themselves up for failure” if they rush into integrating new AI models and shares some advice for businesses adopting the technology.
Listen to the full episode here, along with all previous episodes of the UKTN Podcast.
This season, the UKTN Podcast is speaking to key tech stakeholders in addition to founders of high-growth startups. Tune in every Tuesday for new episodes and don’t forget to subscribe wherever you listen to your podcasts.
The UKTN Podcast is sponsored by Sage.