Can AI help open business?

By Scott Hamilton

We have all been thrown into what the mainstream media likes to a call a new normal. I don’t consider what we are going through with this pandemic to be either new or normal. St. Louis, Mo., was the hardest hit during the 1918 Spanish Flu Pandemic, and many of the same actions helped to stop the deaths in 1918-1919. The necessary actions were not considered normal, just as they are not normal today. Enough about my rant over the “new normal.” What I really want to talk about today is the media hype about Artificial Intelligence and how it can help your business reopen.

AI works on the precept of using the past to predict the future. Amazon uses it to suggest items you might like to purchase based off prior purchases. Walmart does the same with online sales. Google and Facebook target their advertising in the same way. However, things like the COVID-19 pandemic negatively impact these data models. AI does not adapt fast enough to compensate for a society where sudden shifts in food prices and availability cause bulk buying to trend. Manufacturing facilities are closed down and the supply chain begins to fail.

Many experts tout AI like some magic sauce that can be mixed with everyday business process to improve profitability. The problem is, when tragedy strikes, AI fails. The models are based of pre-pandemic behaviors and will predict useless information that can point business in the entirely wrong direction. AI is supposed to be able to replace the white-collar worker, all the way down to delivery drivers and fast food workers. However, it is very clear that it is not progressing as fast as expected and it is far from a cure-all for business woes.

A lot of technology has made large leaps forward during the pandemic. For example, our internet infrastructure has been vastly improved to support the work-from-home force. We have new teleconference companies and great improvements in digital broadcast studio software. However, this is not the case for AI. The hype around AI is subsiding, especially among the companies that are actually using it.

We are beginning to see that AI is far from magical, and actually most useful in accomplishing mundane stuff. We use AI every time we talk to a voice-activated personal assistant like Bixby or Siri and when we use the finger or facial recognition lock on our cell phone. It is used to park cars and in all self-driving vehicles. However, it is mostly used to assist humans in making decisions.

AI was the computing field of the future and now it seems that this particular area is on a downtrend. Uber shut down its AI research lab, Airbnb layoffs included many full-time data scientists, and roughly 6% of all data scientist globally have experienced a layoff since the start of the pandemic. This is a small percentage compared to other industries, but it is extremely high with a field that was thought to be indispensable to large corporations.

I believe what we are seeing is a correction in the AI market, as companies begin to realize that if they simplify their business strategy models, they can eliminate compute hardware costs and costly data science positions to get better analysis in rapid response scenarios. People can make better and faster decisions in rapidly changing environments than a computer is capable. The pandemic is a wake-up call about how poorly implemented most of our AI algorithms are when it comes to real world predictions.