Comprehensive software reviews to make better IT decisions
DataRobot: “Every Business Must Make AI Bias Education a Priority”
DataRobot, a vendor of enterprise artificial intelligence (AI), recently released a report revealing that nearly half (42%) of AI professionals in the US and the UK are “very” or “extremely” concerned about AI bias. The research is based on a survey of more than 350 CIOs, CTOs, VPs, and IT managers involved in AI and machine learning (ML) purchasing decisions in the US and the UK.
Survey participants cite concerns about “compromised brand reputation” and “loss of customer trust” as a result of using biased AI systems. Other concerns include “mismatch with personal ethics,” “loss of employee trust,” and “legal penalty fees.”
Many organizations are already addressing AI bias: 83% of survey respondents said that they had established AI guidelines to ensure “accurate, trusted outputs” and that, in addition, they are doing the following:
- 56% are deploying tools to detect and mitigate biases hidden in training data.
- 60% have created alerts to monitor concept drift (when input data used to predict outcomes diverges from the training data).
- 59% “measure AI decision-making factors.”
Bias prevention initiatives are critically important to organizations – this is confirmed by the survey finding that most of these initiatives are overseen by a CxO, most commonly the CIO (49% of all respondents). And 93% of respondents said that they plan to continue investing in AI bias–prevention initiatives in the next 12 months. Specifically, they plan to invest in the following:
- Model interpretability (59%).
- Internal personnel to manage AI trust (54%).
- Third-party vendors to manage AI trust (49%).
Organizations also leverage external expertise – third-party AI bias experts and consultants – to help them learn about and mitigate AI biases.
Our Take
We are thrilled to see that AI bias is getting the attention of executives and that many organizations are already taking steps to address machine/algorithmic biases. As we mentioned in an earlier note, trust is the foundation of business, and you either have it or you don’t. And given the velocity with which AI and ML are penetrating all aspects of business and society, it is paramount that we build ethical, responsible, and trustworthy AI systems.
And the first step to do that is, as always, education. We agree wholeheartedly with Colin Priest, DataRobot’s VP of AI Strategy, when he says, “There’s more to be done to win the trust of businesses and consumers. Every business must make AI bias education a priority so they can implement critical strategies within their AI systems that will help prevent it from happening.”
Want to Know More?
See DataRobot’s on-demand webinar How to Remove Unfair Bias From Your AI.
To get educated on the various AI biases – from data to algorithms to design and beyond – and to learn about how identify and mitigate them, consult Info-Tech’s blueprint Mitigate Machine Bias.