Quality Assurance for AI

GISKARD is developing an open-source SaaS platform for automated AI quality testing to address ethical biases and prediction errors, aiming to lead in compliance with the EU AI Act.

Subsidie
€ 2.499.999
2023

Projectdetails

Introduction

The adoption of AI technology is growing faster than ever, tripling from 2017 to 2021 (McKinsey). Risks of AI incidents, in particular with ethical biases, prediction errors, and cybersecurity, are rising.

Current Challenges

Current AI quality tools are insufficient and rely on manual testing, leaving a gap that AI/ML engineers cannot meet in terms of workload, costs, and demand at the needed pace.

Proposed Solution

To address this need, GISKARD is developing an open-source and SaaS solution for companies that need quality assurance of their AI models. It provides a software platform for automated AI Quality Testing, Inspection, and Remediation.

Commitment to Standards

As a member of AFNOR, the French national standards council, GISKARD is committed to becoming the leading European software provider to help organisations prepare for the upcoming EU AI Act.

Support and Development

EIC support is crucially needed to achieve this. In this project, GISKARD will optimise and validate its AI Testing solution and extend it to more use cases such as:

  1. Time Series
  2. Computer Vision

Financiële details & Tijdlijn

Financiële details

Subsidiebedrag€ 2.499.999
Totale projectbegroting€ 3.729.395

Tijdlijn

Startdatum1-9-2023
Einddatum31-8-2025
Subsidiejaar2023

Partners & Locaties

Projectpartners

  • GISKARD AIpenvoerder

Land(en)

France

Vergelijkbare projecten binnen EIC Accelerator

EIC Accelerator

A new cardioprotective drug for acute treatment of myocardial infarction

ResoTher aims to validate RTP-026, an immunomodulating therapy, to reduce heart damage and HF risk post-myocardial infarction through Phase II clinical studies.

€ 2.499.999
EIC Accelerator

Clear, scalable and scientific framework to measure terrestrial biodiversity

3Bee leverages IoT, wildlife monitoring, and satellite data to measure and regenerate biodiversity, generating certified Biodiversity Credits for corporations to enhance ESG reporting and brand value.

€ 2.252.714
EIC Accelerator

Novel and Scalable microbial products for REgenerative agriculture

N-Spire aims to revolutionize agriculture by creating a sustainable bioactive fertilizer through innovative manufacturing techniques, enhancing soil health and reducing chemical dependency.

€ 2.499.999
EIC Accelerator

Quantum-based Randomness Processing Units (RPUs) for High-Performance Computation and Data Security

Quside's Randomness Processing Unit (RPU) accelerates stochastic HPC and PQ cryptography by optimizing random workloads, enhancing efficiency and performance across various sectors.

€ 2.499.999

Vergelijkbare projecten uit andere regelingen

ERC POC

Developing Bias Auditing and Mitigation Tools for Self-Assessment of AI Conformity with the EU AI Act through Statistical Matching

Act.AI aims to enhance AI fairness and compliance with the EU AI Act by providing a versatile, plug-and-play tool for continuous bias monitoring across various data types and industries.

€ 150.000
MIT Haalbaarheid

POTENTaiLIZER

Het project onderzoekt de haalbaarheid van een AI-product, POTENTaiLIZER, om MKB-bedrijven te ondersteunen bij het optimaliseren van hun bedrijfsvoering via data-analyse en strategische tips.

€ 20.000
MIT Haalbaarheid

AIQ

Het project onderzoekt hoe de bruikbaarheid van openbare datasets voor AI-training automatisch kan worden beoordeeld om kwaliteitsinzicht te verbeteren.

€ 20.000
MIT R&D Samenwerking

GIFT Generic Image and document Forensics Toolkit

GIFT ontwikkelt een AI-gestuurde toolkit voor document- en beeldforensische analyses om manipulatie en fraude te detecteren, met focus op schaalbaarheid en gebruiksvriendelijkheid.

€ 193.600