On January 10th, Mighty Ai announced a $14 million funding round and its new name. It was previously known as Spare5 and that name will continue on as the name of its AI training community. Why should you care? AI platforms such as IBM Watson provide natural language processing (NLP) out of the box, but still expect you to provide your own training data to teach the system for both language intent recognition and domain expertise. So, whether you are building your own AI platform or using a third party for your application, you need training data. It is a building block of AI systems.
Providing Easy Access to Training Data Sets
Mighty Ai provides what it calls Training Data as a Service (TDaaS). The need for training data is one of the key reasons that many people speculate that the AI technology wave will be dominated by the big technology players. The big companies already have vast troves of training data and access to a large customer base that can continuously improve the data sets. Mighty Ai could help fill this need and make AI innovations far more accessible to startups. The company intends to do this by making its current training data available along with access to its training community, Spare5.
Some large players that don’t have their own training data sets or AI platforms see this as an opportunity as well. Intel Capital and Accenture Ventures participated in the round along with blue chip venture capital firms, GV, Foundry, Madrona and NEA. Partnerships with Intel and Accenture were also announced.
Mycroft, an AI-driven voice application platform competing with Alexa and Google Assistant, is using the open source community to accumulate its training data. That is not a route available to all startups. Mighty Ai is a commercial alternative. Today, you go to Amazon Web Services for instant availability of computing power. Tomorrow, you may go to Mighty AI for instant access to training data and a training community for your AI application. Or, if you are building on Mycroft’s platform you reach out to its community.
Expect More Vertical and Horizontal TDaaS Solutions
This may turn out to be an important announcement for the AI startup landscape. If the model works, you should expect to see more TDaaS companies and eventually a variety of vertical and horizontal domain specializations such as healthcare, environmental science, visual recognition and many more.