A.I. strikes back

deep thinking ai artificial intelligence
Credit: Thinkstock

Is your new artificial assistant reading this blog?

The term artificial intelligence (A.I.) was first coined by John McCarthy. McCarthy was teaching mathematics at Dartmouth at the time. The year was 1956. Now, 60 years later, we are at the dawn of a new age of artificial intelligence. The first age passed without bringing the world domination by super human robots that we saw in the movies. This time, things could be different.

McCarthy had a saying about artificial intelligence: “As soon as it works, no one calls it A.I. anymore.” This is certainly true today. We have A.I. all around us. Amazon is always sending me emails about things it thinks I might like. Alexa, the Amazon assistant I have at home, answers my arcane questions and even turns my lights on and off.

Self-driving cars aren’t just driving around Google's headquarters, they're also driving around Pittsburgh. In IT, our security is strengthened thanks to A.I.

There's a new arms race to create the most intelligent assistant. Siri, Alexa, Cortana, Amelia, Watson — the list goes on and on. Google Photos can search for pictures of my family, or even for pictures of a birthday cake. Artificial intelligence is here, but the key thing to know is that it’s about to get a lot smarter and a lot faster.

One of the hottest areas in A.I. today is the effort to develop self-driving cars. One of the most talked about companies in the self-driving car race is Tesla. Let’s look at what enables Tesla to progress toward the dream of the driverless car.

  • Fleet learning: One of Tesla’s unique advantages is the fact that its cars are all connected together in a way that enables them to learn from one another. The data they collect while driving can help improve the quality of the digital map, and enable each car to learn based on the data collected by all of the other vehicles in the fleet.
  • Computing power: NVIDIA has created an A.I. “car computer” that consumes just 10 watts of power. The ability to do edge computation (in the vehicle) in addition to central “cloud computation” is key to delivering advanced A.I. systems like self-driving cars.
  • Sensors and big data: Cameras, radar and ultrasonic sensors are cheap enough to be used in mass-produced vehicles. Similarly, barometers, cameras and accelerometers are in most modern smartphones. These sensors give A.I. systems data they need to learn.
  • Agile development: Tesla is able to update its cars easily over the air. The algorithms can be continually improved because the software is built for continuous deployment.
tesla autopilot model s large Tesla

A Tesla Model S with Autopilot engaged.

Now let’s generalize these enablers to business in general. There are five things that are needed for A.I. to really take off, and they're all happening right now.

  • Cloud and computing costs: High computational speeds that are affordable.
    • The cloud provides a huge pool of computing resources with which to do machine learning. The cloud also enables things like fleet learning, where all systems benefit from what others have learned.
    • Edge computing: Many people now have a supercomputer in their pocket because it’s economically feasible. The combination of a supercomputer at the edge talking to the cloud enables whole new capabilities.
  • Big data: Massive amounts of data for learning.
    • The ready availability of large amounts of data is one of the developments that has advanced machine learning. Image recognition is a good example: It has advanced so fast because there is a huge library of images to train on thanks to those cameras on everyone’s smartphones.
  • Sensors and distributed intelligence nodes: Smartphones.
    • You carry an array of sensors in your pocket that contribute large amounts of data for machine learning. Think about data for healthcare (heart rate, steps, gait) or weather (barometer, temperature).
    • Your smartphone enables you to carry intelligence with you in your pocket to perform local computations.
  • Natural language processing: Watson, Siri, Google, Cortana, Alexa…
    • Everyone can use it.  It doesn’t require technology skills.
  • Continuous improvement software development.
    • Agile and continuous improvement development techniques allow learnings to evolve quickly, accelerating improvement to A.I. systems.

The return of A.I.

So, artificial intelligence is back. My master’s thesis was an expert system. Now, almost 30 years later, A.I. technology is readily available to millions of people who will enable even more amazing things to come to light.

Of course, as soon as it works, no one will call it A.I. So start developing your strategy for A.I., or A.I. might develop a strategy to replace your business first.

This article is published as part of the IDG Contributor Network. Want to Join?

Drexel and CIO.com announce Analytics 50 award winners
View Comments
Join the discussion
Be the first to comment on this article. Our Commenting Policies