- Search Solutions
- Leadership, Assessment
- Career Transition
- About Optimum Talent
- Start Your Transition
If your employer has provided you with Career Transition Support through Optimum Talent, click here to get started.
Written by: Jocelyn Bérard | President, Central Region | Optimum Talent
Before COVID-19 swept across the globe, there were very high hopes that Artificial Intelligence and Machine Learning would help governments, researchers and health care providers learn from past events and behaviour to help plot a response. It did not take long for those hopes to be dashed.
Machine-learning models are designed to analyze data – huge tracts of data – to identify patterns and make predictions about the future. But they are only as good as the data they are able to incorporate. If the data is flawed, or if there is a huge lurch in behaviour that undermines historical data, then machine learning can be undermined.
Take online retail as an example. A recent article in the MIT Technology Review reported that within a few days of COVID-19 being declared a global pandemic, the top 10 most popular products sold on Amazon changed completely.
Gone were the phone cases, chargers and Lego that had served as the backbone of Amazon sales. In their place, people were buying toilet paper, face masks, groceries, and disinfecting products. By February, the top-10 seller lists in every major Amazon category were – one way or the other – connected to COVID-19.
The MIT article noted that AI is only as good as its data, and when subjected to seismic change, so too are machine learning’s predictive powers. “Machine-learning models are designed to respond to changes,” wrote Will Douglas Heaven, a senior editor at MIT Technology Review. “But most are also fragile; they perform badly when input data differs too much from the data they were trained on.”
Read more on the topic of technology here!