Deasie's platform scans files for sensitive information before they are incorporated into a training dataset and ranks the remaining records by relevance. This ensures that only high-quality data is used for AI training. The platform can also automate the process of moving training data spread across different systems into a centralized repository. The company revealed that a multibillion-dollar enterprise has signed up to pilot its platform and that its customer pipeline includes 30 other companies, five of which are in the Fortune 500.
Key takeaways:
- Lakera AI AG and Deasie Inc., two startups focusing on reliability in large language models, have raised funding to increase the adoption of their products. Lakera secured a $10 million investment, while Deasie raised $2.9 million in seed funding.
- Lakera's main product, Lakera Guard, is a cloud service that can detect attempts to trick an AI into generating harmful output or extracting sensitive data. The company also offers MLTest, a software designed to help developers test their computer vision models before releasing them to production.
- Deasie offers a platform that prevents unreliable information from entering a large language model’s training dataset. The platform scans files for sensitive information and ranks the remaining records by relevance to ensure only high-quality data is used for AI training.
- Deasie has reportedly signed up a multibillion-dollar enterprise to pilot its platform and claims to have a customer pipeline that includes 30 other companies, five of which are in the Fortune 500.