COVID-19 also shakes up machine learning models

While machine learning is great to automatize certain tasks, there are also various weaknesses. For example, when conditions change significantly, it's possible that models that were trained using older datasets can no longer provide the required results.

The Register writes that's the case today with the COVID-19 pandemic. Citing a report from Gartner, the site reports that the extreme disruption in the aftermath of COVID-19 has invalidated many AI models that are based on historical data.

Human behavior has changed and this means machine learning models for product recommendation engines or next-best-offer type of recommendations no longer work as before as customers are behaving very differently:

In any case the 'new normal' is still emerging, making the validity of prediction models a challenge, said Rita Sallam, distinguished research vice president at Gartner.

"It's a lot harder to just say those models based on typical data that happened prior to the COVID-19 outbreak, or even data that happened during the pandemic, will be valid. Essentially what we're seeing is [a] complete shift in many ways in customer expectations, in their buying patterns. Old processing, products, customer needs and wants, and even business models are being replaced. Organisations have to replace them at a pace that is just unprecedented," she said.