When crowdsourced labor company CrowdFlower recently raised funding from Microsoft, co-founder Lukas Biewald told me his team was focused on technologythat allows businesses to supplementalgorithms and artificial intelligence with human judgment from crowdsourced labor pools.

Now CrowdFlowerbringing on more experts to shape the development of that technology. Specifically its formed a three-person scientific advisory board, made up of Barney Pell (founder/co-founder of startups including Powerset, LocoMobi and Moon Express, who also led an artificial intelligenceteam at NASA), Anthony Goldbloom (founder and CEO of Kaggle) and Pete Warden (a staff research engineer at Google, where hes the technical lead on the TensorFlow Mobile machine learning project).

With all these different customers and all these different applications, we wanted them to be confident that theyre going to get a high-quality algorithm, said Biewald. (He was previously CrowdFlowers CEO and now serves as its chief data scientistand executive chairman. Hes also a friend of mine from college although we really only talk about CrowdFlower now, which is kinda sad when you think about it.) One way to make sure all the product decisions we make really reflect the cutting edge was to get some ofthe world leaders come in and look at our product.

Pell, who will be co-chairing the advisory board with Biewald, has a long history with CrowdFlower hes already an investor in the company, and he noted that Biewald first came up with the idea while working at Powerset. He said CrowdFlowers human in the loop approach, where humans can help provide training and quality control to AI, could become increasingly important.

When peoplethink about AI, theyre generally thinking about 100 percent automated solutions,Pell said. But the reality is, If theres people in the loop somewhere, then where youre really confident, [the algorithm] can handle those cases, and then the rest of the marginal cases go to people.

Pell added thatCrowdFlower (which launched at the TechCrunch50 conference) works with customerswhose technology might seem fully automated at first, such as self-driving cars but even in that case, they still need humans to help with train the vision systems and help with the labeling.

As for the boards role, Pell said it will both look at individual products under development and at the broader CrowdFlower roadmap.

I brought up another possible benefit: In an industry where artificial intelligenceand machine learning have become buzzwords thrown around by every startup, thiskind of boardcan add an important layer of credibility.

The real challenge here for any company thats trying to do machine learning is that theres so much research that its impossible for anybody to synthesize it all, Biewald replied. I think youll see more and more companies trying to adopt an approach like this.

Read more: