Mistral’s new chatbot: Unpacking Le Chat’s Capabilities and Controversies

Author(s): Jelle Van Bost

Originally published on Towards AI.

Testing the Limits of Mistral AI’s Conversational Assistant in Beta Access
Image generated by the author on firefly.adobe.com

Mistral's new AI tool promises powerful conversations, but after a few simple tests, I discovered that it might have access to more than just your words. This morning, I received an email from Mistral AI inviting me to test their new groundbreaking conversational assistant. A misspelling in the subject line made me chuckle, but could this newcomer rival the power of existing chatbots like chatGPT and Gemini?

Screenshot by the author.

I am a self-proclaimed LLM enthusiast, always trying to push the boundaries of these tools. In this article, I’ll give a quick overview of the new Mistral AI’s “Le Chat”. I’ll test their capabilities with three simple questions designed to test their reasoning, writing and coding skills.

Of course, no AI is perfect. I’ll also try a few tricks to see how well these models are protected against potential misuse.

Disclaimer: This is a beta version of Le Chat. This means that the output might be improved with the final release. Every opinion expressed in this piece reflects only my personal opinion.

Before diving into Mistral AI’s capabilities, I hit a major red flag — Their terms and conditions. Take a look at the following section:

Screenshot by the… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI