We have read, heard, and discussed a lot about the influence of artificial intelligence on our lives. The advent of AI and machine learning has actively brought changes in the way people interacted with each other, in the way business used to be conducted, in the way people listened to their music or chose a TV show to watch after dinner.
The COVID-19 pandemic has turned the tables. Wearing masks, sanitizing your hands every now and then, staying home as much as possible, staying anxious about a virus, this is our reality, the new normal. AI depends on reiterative patterns in our behavior. Our behavior has been radically affected by the current health crises; let us look at its impact on AI and machine learning initiatives.
Looking for AI in the industry
The application of artificial intelligence as an autonomous system as well as an augmentative mechanism has spread across many industries. Be it creating a marketing strategy, or detecting fraudulent transactions, AI has its uses. Regardless of what task its performing AI depends on training machines with data. For instance, when you browse through your entertainment system, say, Netflix, a bot tracks your viewing history, popular trends, the time you spent on a certain sort of content and suggests you the next movie. Similarly, it helps the producers to understand what they need to do to keep you happy. We can see that it really counts on our habits, a logical flow in our thought process which is reflected by our activities.
Disruption of the disrupters
In a matter of 5 days in late February this year, COVID-19 related products knocked all other products down from the list of top ten best sellers on Amazon. This has caused hiccups in inventory management systems across industries. While a lot of automation systems are coping, some have stopped working as they should have. Machine learning models are meant to adapt to a different situation but it runs the risk of performing badly when the data drifts too far apart from the data it was trained on. For instance, an Indian manufacturer of sauces depends on machine learning to manage inventory. The spike in sales has taken the system by surprise and its analyses have become inaccurate and impractical. Similarly, a popular streaming house is having issues with the AI-based recommendation system.
There is a coordinative relationship between AI and human behavior and it is time we addressed this fact properly. A lot of companies suffer because although they buy machine learning systems, they do not have the in-house capacity to manage it.
The human touch
Just as the speculations regarding complete automation or the irrelevance of human workers had started to take flight, the pandemic taught us not to get ahead of ourselves. A lot of companies are counting on the occasional human rectification so that their automotive systems can cope with the changes. For instance, a British company Phrasee uses AI to produce phrases for email or Facebook marketing campaigns. But they also use human eyes to check them because AI-based recommendations can go wrong. Right now, they have completely abandoned phrases that indicate excitement, happiness, or such phrases that may induce anxiety.
It is a difficult time for humans and also an opportunity to understand the synergic existence of AI and human beings.