AI bias means these tools can only act on what they know. So when an anomalous event happens, like the global covid pandemic in 2020, people have to intervene and adjust the models used by the AI. Here’s how the pandemic readied Alibaba’s AI for the world’s biggest shopping day.
Alibaba and JD invested heavily in AI to predict shopping demand, optimize global distribution and streamline worldwide delivery. The pandemic made AI less able to predict behaviour, so Alibaba reviewed its short-term forecasting strategy, factoring in more immediate variables like the previous week of sales leading up to major promotional events, or external data like the number of covid cases in each province.
As live-streaming ecommerce exploded in popularity during the periods of lockdown in the pandemic, the company’s logistics arm also built a new forecasting model to project what happens when popular live-stream influencers market different products.
Find out more, from MIT…
Implications for AI project
At the start of an AI project, invest time in considering the types of bias you could be introducing through your choice of data.
