Four Predictions for Artificial Intelligence in 2017

2017 holds a lot of opportunity in artificial intelligence, particularly for investors.

2016 is the year artificial intelligence went mainstream.

And I don’t just mean the naughty photos filter on Twitter or the fake news suggestions on Facebook.

Fueled by unprecedented funding (and a growing open source ecosystem), founders are delivering artificial intelligence startups at a record rate.

  • Apple, Facebook, GE, Google, Intel, Microsoft, Salesforce and Samsung–and more–made major AI investments last year.
  • Five million homes are now talking their music and shopping choices through with none other than Amazon Alexa.
  • Self-driving cars have their own new U.S. Department of Transportation Committee. A few years ago, people were talking about 2025 or so for self driving (level 5 autonomy)–now it’s before 2020. It’s compelling that self-driving may whittle down our 1.2 million annual auto deaths.
  • Not one but two AI unicorns grew their horns — iCarbonX in China and Cylance in Silicon Valley.
  • Over one fifth of the MIT 50 “smartest companies” list have AI as a core approach.Bringing AI home

So in the midst of this hype, how do figure out how this affects you? Here are some predictions on where AI is heading in 2017 to help you bring some of the hyperbole down home.

1. Jobs vs. artificial intelligence

There’s no doubt that artificial intelligence approaches can eliminate some sectors of jobs. Port Botany in Sydney already has a “human free” zone where thousands of shipping cartons are managed by AI-informed robots, for example.

The truth is, AI also creates jobs.

Accenture research calls artificial intelligence the “new economic superpower.” Rather than eliminating jobs, their indications point the overall impact of AI as boosting productivity up to 40 percent and doubling the annual growth rate of developed countries.

For next year, I predict that the increasing need for humans-in-the-loop is going to more than keep up with any jobs eliminated in the near term. People will be working with AI for things like:

  • validating the results of artificial intelligence
  • preparing the data for machine learning, and
  • creating and designing systems.
    Not only that, but also the need for more data science, as well as updated regulations and ethics, will create dozens of categories of interesting new jobs.
  • All of these categories create opportunities for entrepreneurs and investors alike.

2. Artificial intelligence funding: (even more) momentum

Artificial intelligence exits topped a billion last year for the first time. In the year ahead, expect that pace of acquisitions to heat up even more as companies that aren’t the usual suspects get more comfortable with machine learning, NLP and the entire AI approach. Industries outside of traditional tech–like call centers, shipping centers, and warehouses–will be buying artificial intelligence. 2017 will also see more mega validation like Stackpath, the Texas security-as-a-service-with-AI startup that raised $180 million in its series A. It’s a good time to be investing in AI.

3. Artificial intelligence acquisitions: record deals ahead

The deal frontier will see more, more and more–more numbers of deals, more record big deals, and more types of deals, including deals from corporate venture investors.

  1. In terms of major acquisitions, I’m expecting early wins from the developer toolsets that build on open source platforms.
  2. Plus, we should see lots of acquisitions in front-end communications, like chatbot customer communications tools.
  3. One of the most exciting and acquirable emerging frontiers in AI is that between data and the developer. For example, last year, former Skytree co-founder and former Georgia Tech associate professor Alexander Gray made his machine learning platform open source. There are many more, including open platforms from Google and Microsoft. Cynthia Harvey recently posted a great list of open source AI tools. With this much developer opportunity on hand–literally–smart developers tend to develop some smart tools for themselves, especially around the thorny knots of data cleansing for machine learning.

4. AI and the media: hype, hope and hiccups

If last year was the year when the AI hype caught media attention, 2017 is set to be the year when AI hiccups do. From cyber fraud to shopping, as AI is implemented its vulnerabilities are going to be exploited by kids and criminals alike. One blip on the radar last year happened when the BBC reported on a medical AI that misclassified asthma patients as low risk for pneumonia death. That’s because people who have asthma and pneumonia often go straight to intensive care–because of, and not in spite of, the mortality risk. The AI bloopers reel is just beginning and will start getting good this year. I predict learning to laugh at the profound stupidity of machines is a skill we all get work on before 2017 is over.

First published in my innovation column in Inc. Magazine.