AI21 Studio

Conversational AI Language Model Startup AI21 Labs Raises $64M

Natural language processing (NLP) AI startup AI21 Labs has raised $64 million in a Series B funding round led by Ahren. AI21 Labs offers an array of tools to build custom conversational AI assistants based on its Jurassic-X Jumbo language model. The new funding bumps AI21 Labs’ valuation to $664 million and will go to scaling the company as it jostles with large-scale language models like OpenAI’s GPT-3 and Google DeepMind’s AlphaCode.

AI21 Labs

AI21 Labs is known for creating AI writing assistant Wordtune, a browser extension that helps transform ideas into concrete language for written communication and its cousin Wordtune Read for document summary and analysis. More recently it created AI21 Studio to leverage its NLP technology for creating custom pre-trained language models for conversational AI chatbots and voice assistants used by businesses including grocery chain Publicis. The company recently drew attention after releasing a chatbot trained on the texts of late Supreme Court Justice Ruth Bader Ginsburg’s legal opinions. AI21 Labs described the chatbot as a flawed experiment that might respond in ways Ginsburg never would, but it served to raise awareness of the startup and some of the possibilities of its platform. The new funding round nearly doubles AI21’s total investment to $118 million.

“We completed this round during a period of market uncertainty, which highlights the confidence our investors have in AI21’s vision to change the way people consume and produce information,” AI21 Labs co-CEO Ori Goshen said. “The funding will allow us to accelerate the company’s global growth, while continuing to develop advanced technology in the field of natural language processing. We are looking forward to growing our team and our offerings.”


OpenAI Debuts New GPT-3 Model to ‘Align’ With Human Intent

DeepMind Claims New AlphaCode AI Programmer Outperforms GPT-3 and is Competitive With Humans

Microsoft Enhances Azure OpenAI Service and Widens Access to GPT-3 Models