Artificial Intelligence is the Future

By Wendy Gonzalez,Managing Director, Samasource

The growing demand for AI and Machine Learning is evidenced by the 950 plus AI companies that have flooded the market globally, representing nearly $5 billion in funding, according to Venture Scanner.

AI is powered by an area of computer science called Machine Learning, which is the ability for a software program to self-learn through pattern recognition.

It is a way of programming that manages everything from computer vision so your Wii can follow your every dance move and score you correctly, to contextualized user experience so Netflix can make smart, personalized movie recommendations. Artificial intelligence (AI) is no longer reserved for movie effects and science fiction, but an area of significant growth and investment for businesses.

In fact, Gartner listed Machine Learning as one of the top trends for organizations to pay attention to in its 2015 Hype Cycle. It’s bigger than Big Data and also considered one of the primary tools that companies will use to gain a competitive advantage in 2016.

This number accounts for about 5 percent of total VC investments for 2015, which is still on the lower side but with 40 percent of the world’s population now online, we’re creating more data assets--material that describes our behaviors, interests, and knowledge--that will continue to propel the category forward.

Machine Learning algorithms are programmed to get smarter and smarter as they recognize patterns to the point that they are able to automate end-to-end processes or perform tasks like facial recognition. To recognize these patterns, algorithms need to be trained, and this training for artificial intelligence is powered by human intelligence, not machines.

BPOs and crowdsourcing platforms are training the most futuristic, cutting edge products and personalized services by making simple human judgments, hundreds and thousands at a time. The training starts broad until the algorithms are fine tuned, focusing on human judgment for nuanced scenarios or validation.

In the case of Samasource, we “Impact Source” work related to some of Silicon Valley’s boldest projects. For example, we partnered with @WalmartLabs to take their algorithm for predicting and assigning gender (e.g. so “Nike Women's Running Shoes” appear when you categorize items for women) from 59 percent to 93 percent accuracy by manually QA-ing (correcting or assigning the proper algorithm) 115,500 products.

This work can also come in the most unlikely of forms. We recently started working with Paul Allen to promote sustainability through annotating elephant images taken from hours of drone footage intended to automate detection of poachers for the Great Elephant Census.

Every time I visit our Samasource center in Nairobi and watch as nearly 400 workers pour through mounds of data, I can help but become acutely aware that machines will only take us so far. It’s humans that are powering our world’s most disruptive and innovative technologies.