There's is a vast ecosystem of free and open source projects that offer access to statistical and machine learning techniques now usable by non-specialists. I c

There's is a vast ecosystem of free and open source projects that offer access to statistical and machine learning techniques now usable by non-specialists. I can now easily build recommendation systems scaling to billions of ratings, do facial recognition with a high degree of accuracy, fraud detection, or identify all the things in a photo (thanks ImageNet!). I can build models to predict customer churn or ad CTR performance, extract summaries from large text documents, use unsupervised clustering to identify potential customer segments, determine if an image contains nudity, or if a 140 character tweet contains positive or negative language. Open access has come a long way in a short period of time. It was only five years ago that Facebook introduced auto-tagging friends in photos (facial recognition), a feature that's now taken for granted but was once the realm of bad CSI TV shows and Las Vegas casinos. Google, Facebook, Amazon, Microsoft, and IBM even started a consortium this September to share research [0]. "AI has been democratized" might be a bit overzealous, but I'm perfectly happy with the offerings on the table right now. [0] https://www.partnershiponai.org/ https://www.partnershiponai.org/