At the end of 2018, I sat down and, wearing my Nostradamus hat, speculated on what major AI trends were in store for the year to come. In 2019, much has been done and written about AI — from doomsday scenarios (loss of jobs, privacy) to AI built to win games of Go, poker and StarCraft to the coming of artificial general intelligence.
But aside from the hype, what predictions came true? What have been the real, tangible and useful applications of AI in 2019? And what’s coming in 2020?
Here’s the rundown:
I believed that 2019 would be the year of specialized AI systems built by organizations based on their own data.
This prediction has definitely come true. The idea that you can just “download AI” to solve a real-world problem is a huge misconception. Rather, AI needs to be specialized to the unique datasets that enterprises own, understand and, ultimately, use. Enabling enterprises to quickly build, deploy and iterate on custom AI applications has been the key for 2019 and will be the central theme in the year ahead.
In 2020, user-friendly and affordable AI platforms will be critical to enabling enterprises to mine their own data to build customized AI solutions, making AI more applicable to a much broader range of applications. As enterprises learn how to put AI to work for them, the understanding and usefulness of AI will grow significantly, leaving the doomsday predictions and theoretical musings far behind.
In 2019, I expected that the industry would come together to foster a global economy around AI.
This is where I was wrong. I imagined a 2019 where industry leaders and nations would come together and realize that AI has the potential to help solve some of the world’s most pressing challenges and benefit society in a number of ways. I envisioned a 2019 where, as a society, we opened the door to conversations between the many nations leading the AI revolution.
The opposite has occurred. 2019 has seen polarization as a result of AI superpowers taking shape. And rather than coming together, large corporations have begun amassing and mining personal information with dollar signs in mind rather than global interest.
While AI and its architects (scientists and engineers) have kept their promise of building more useful tech, not everybody around the AI table has ethics top of mind. Hopefully, like-minded AI companies will come together in 2020 to establish a code of conduct for AI that’s guided by ethics and human-centric thinking. I believe it can be done.
Looking ahead, AI needs to live and learn at the edge.
Most AI systems today are born under the premise that learning occurs in centralized cloud locations, forcing companies to store and potentially even renounce ownership of all their data. And as a corollary, the only way organizations can access their data is by paying large cloud providers massive fees.
Well, 2019 has shown us that many enterprises are saying no to giving up their data, having to ping the cloud and adopting this modus operandi altogether.
In 2020, customers will demand AI that can be trained, deployed and refined at the edge so that processing occurs where data is generated, minimizing latency problems, privacy issues and massive cloud fees. This is particularly true for industrial and manufacturing applications. More and more manufacturers will turn to the edge in order to maintain ownership of their data and reduce dependence on the cloud. As a result, reliance on the cloud will be limited to certain use cases or particular scenarios.
The good news is that as we move into 2020, more and more enterprises are attempting to deploy AI as a part of their business strategies. The bad news is that many have failed or have failed to achieve the metrics they wanted.
My hope for 2020 is for there to be less of an effort around frivolous applications of AI and more emphasis on smaller, more focused but real applications of AI that deliver clear benefits and return on investment. Perhaps without the fanfare of AI beating humans at the next game, enterprises can develop applications that save money, improve efficiency and improve our quality of life.
