Startup That Uses Humans to Train AI Just Nabbed $18M in Investments

Do you think artificial intelligence (AI) is becoming too autonomous and independent of humans? Well, think again! Scale is a startup that utilizes a team of contractors to train AI systems by categorizing visual data.

It’s making big waves in AI development — in fact, it just raised an additional $18 million in investments.

Not Your Usual AI Startup

According to Pitchbook, this brings Scale’s valuation to a total of $93.5 million. The investment comes on the heels of a successful two-year growth period for the San Francisco development company. In just the last year, revenues have increased by 15-fold (although specific numbers weren’t revealed).

Scale first debuted in July 2016 and presented itself as a smarter Mechanical Turk. Basically, it aims to address more complex demands of AI systems that are a little too much for what Turkers usually do. As CEO and Co-Founder Alexandr Wang puts it, “We’re honing in on AI broadly. Our goal is to be a pickaxe in the AI goldrush.”

The company covers a wide range of services, such as transcription, categorization, comparison, and even phone calling. But recently, a large portion of business has come from self-driving car companies, specifically for image recognition.

A Specialty in Self-Driving

A human is needed to help train AI systems on subtle but important details, like the difference between a gray-colored piece of clothing and a piece of asphalt. Wang says, “This sub-segment of AI, autonomous vehicles, really took off after we launched, and that segment has been the killer use case for us.”

Scale has helped the likes of General Motors, Lyft, Voyage, and Nuro develop autonomous vehicle systems. Organizations like these tap into Scale’s armada of around 10,000 contractors to help sort raw data through services such as Sensor Fusion, Image Annotation, and Semantic Segmentation. Scale says it has helped annotate 200,000 “miles of data” from self-driving cars.

Of course, Scale still helps out in several non-automotive initiatives. For instance, it assisted Pinterest and Airbnb in building out an AI-powered recommendation system and visual search, respectively. But it’s hard to ignore the demand that self-driving car companies have. Especially when you’re one of the only AI companies equipped to handle a myriad of different problems.

The Aim of the Investment

The latest funding round for Scale was led by Index. With existing investors like Y Combinator and new ones like Dropbox CEO Drew Houston participating, the company managed to raise $18 million in its Series B round. Scale already has big plans for the new investments.

Of course, a substantial portion of it will go towards improving existing services, especially in the area of visual data analysis. “Our first goal is to improve algorithms for customers today,” says Wang. “There is no limit to how accurate they want to make their systems, and they need to be constantly feeding their AI with more data. All of our customers have this, and it’s an evergreen problem.”

The second goal is to expand beyond self-driving and visual data sets. “Right now, so much of the success has been in processing imagery and robotics or other perception challenges, but we really want to be the fabric of the AI world for new applications, including text or audio. That is another use of funds to expand to those areas,” explains Wang.

It will be interesting to see how successful Scale is at this second part. But one thing’s clear — they’re already leagues ahead of the competition in AI annotation for self-driving cars. And that’s already enough to secure them a bright future in AI development.