The short answer is that, yes, there is a risk of bias when it comes to artificial intelligence in business. 42% of AI professionals across both the United States and United Kingdom said that they are concerned about AI bias.
This survey was held in the wake of a study that found that 71% of organizations currently rely on AI in order to execute up to 19 different business functions. There are plenty more that use AI to execute up to 50 different business functions.
What Do Businesses Use AI for?
AI being used by organizations is there to execute functions across a number of different departments. This includes finance and accounting, operations, sales, marketing, and human resources. The biggest concerns about bias among IT executives that were surveyed were loss of customer trust and compromised brand reputation.
Some of the biggest challenges in developing unbiased AI algorithms are in determining what data to train AI models with as well as understanding how to input data that relates to AI decisions.
Learning More About AI
One important step in the process is that IT decision-makers are using tools to learn more about why AI makes negative decisions. That includes using tools to check which input data has the biggest effect on an AI decision. Things such as word clouds are used to look at how text input works in association with AI decisions.
To enhance bias prevention efforts, IT experts are planning to invest in more sophisticated white box systems; these are where AI decisions are explainable. Plenty also said that they will hire internal personnel in order to manage AI trust while others will bring on third-party support.
Ultimately, methods are being implemented to combat AI bias but there is still a heavy reliance on AI by businesses across the US and UK. How that will change over the coming years remains to be seen and will likely see even more developments made as the years move forward.