Why is AI making a comeback?
This is Part 3 of a 4-part series on “4 Questions About AI You’ve Always Wanted to Ask". The next post is “Part 4: Why do businesses want AI? What are they using it for?”
AI research has roots back to 1954 and received big funding in the 1960’s due to interest from the US Department of Defense. Like all technologies, AI has had hype cycles with fluctuating levels of interest and funding. There are three basic reasons why AI has made a comeback recently.
First, there is a whole lot more data in the world now. We live our lives as digital, social media beings and that leaves a great variety of rich data behind as a by-product. Even the somewhat antiquated healthcare system has finally seen a transition to the digital age with Electronic Medical Records of patients plus new collective doctor diagnostic systems which allow for sharing of data amongst doctors – when agreed by patients. There is new genomic data thanks to the popularity of genetic testing for ancestral heritage.
Social media data and geo-tracking data are growing exponentially thanks to Google, Facebook, You Tube, Twitter, What’s App, LinkedIn, Instagram and Snap Chat. The Internet of Things or connected devices are creating billions of bytes of data doing everything from detecting maintenance needs in oil rigs to ensuring your house stays safe, sound and a cozy temperature.
Retailers are collecting buyer behavior data from apps you have downloaded for convenience or rewards like Starbucks, Amazon, Walmart and Target. There are slews of services apps that you are using that are creating data profiles as well such as: banking, matchmaking, ovulation tracking, apartment/house finding/evaluating, financial advisement, gaming. Just look at your smart phone right now. All of those apps are collecting data about you either directly or from geo-tracking on your phone or pictures you have taken or all of the above.
Not to mention companies are monetizing their existing data or augmenting their data to develop data sets specifically for use with potential business partners’ AI systems. For example, fitness tracker groups are combining their data with other health data via Application Programming Interfaces (APIs) to allow more insight into healthy habits like working out plus nutrition for weight loss goals. While another group is tagging pictures of specific crops at different levels of ripeness so that farmers who own specific robotic harvesting tractors can delineate right there in the field which ones are spoiled and which crops are good.
In this new AI era, data and the training of the algorithms is where the differentiation will come not the algorithms themselves. Hence, why I spend so much time explaining all the possible data sources so you can see for yourselves how diverse and big the data opportunities are in the AI space.
Second, the computing power is exponential with new quantum capabilities paired with data storage - which has gotten more efficient and down-right cheap. For the first time in history, computing power, the ability to crunch through all of the data is actually keeping up with ambitions for AI systems. Though those who train the algorithms will complain that it is still taking weeks instead of hours, but the computing power is evolving quickly. Fun fact, when IBM’s Watson first came on the scene with Jeopardy back in 2010, the computing hardware took up a huge room and could do a lot less crunching at the same time. As of 2017, the same tech was the size of three pizza boxes and it just continues to go down with the introduction of each new chip.
Third, more businesses are jumping in with more applications of AI than ever before, thanks in no small part to the large variety of data available now that wasn’t available in the past. Government armed forces and intelligence groups have always been interested in AI. Banks have had need for machine learning aspects of AI as long as hackers and fraud have been around. But more recently insurance firms, manufacturing groups, large technology companies, media and entertainment firms, oil and gas companies, retailers, automotive companies, have all come up with inventive ways to leverage AI in their businesses. IBM had a log of over 150 different use cases in almost every single industry on the planet; and that was back in 2015.
Next week's article will be: "Why do businesses want AI? What are they using it for?"
As always, if you have any questions, please contact, Cortnie@AITruth.org .