Helping keep quality in AI industry job one, Giskard raises €1.5 million

Parisian software startup specialising in artificial intelligence quality, Giskard has raised €1.5 million. The investment will provide the startup with the firepower to engage in a sustained recruitment drive that will welcome additional talent to develop and commercialise its offer while venturing into new fields including computer vision, time series, and generative AI models.

Giskard’s €1.5 million round was led by Elaia, with participation from investors including Bessemer Venture Partners and well-known angels including Christine Balagué (Founder of Good in Tech), Julien Chaumond (CTO of Hugging Face), Oscar Salazar (Founding CTO of Uber), Nathalie Gaveau (Co-Founder of PriceMinister), Charles Gorintin (CTO of Alan), Nick Hernandez (CEO of 360Learning), Vincent Huguet (CEO of Malt), Shéhérazade Semsar de Boisséson (CEO of McCourt Global), Chris Schagen (ex-CMO of Contentful) and Victor Coisne (VP Marketing of Strapi).

Founded by Alex Combessie (CEO), Jean-Marie John-Mathews (CPO & AI Ethics researcher), and Andrei Avtomonov (CTO), three experts previously working on business-critical AI systems at Dataiku & Thales, Giskard is taking on a staggering statistic: Since 2019, over 2,000 incidents related to Artificial Intelligence (AI) models have been reported on the AI Incident Database.

Needless to say, these “incidents” can, and have, posed economic & ethical risks for businesses and the general public, often leading to varying degrees of mistrust.
 
“Giskard defends the vision of a Responsible AI that serves the business performance of companies & respects the rights of citizens. To embody this mission, Giskard enables AI professionals to collaborate with business teams to evaluate and test the quality of AI models,” says Alex Combessie, co-founder & CEO of Giskard.

The company’s AI Inspect product helps visualise and explain AI models in order to identify errors and biases, thereby making the AI development process more efficient, saving time, and greasing the wheels, as it were,  between AI and business teams.

Adding to the upping of the quality, Giskard’s AI Test product allows for the building of automated tests on AI models to measure critical quality criteria, including ethics, robustness, and efficiency. And here lies the crux of the argument against AI as these tests are designed precisely to prevent the risks of costly AI “incidents”.

On the investment Elaia partner Sébastien Lefebvre commented, "As AI becomes more ubiquitous, the more impact it will have on our lives. This increases the need for testing models and ensures no detrimental biases to any party. This is why we’re excited by the Giskard team and their ability to be besides data scientists who are building the future of AI.”