SG Orgs Struggle with Data Bias In AI
- By CDOTrends editors
- April 26, 2023
A recent global survey by Progress, a trusted application development and infrastructure software provider, highlights data bias as a significant challenge for 68% of Singaporean organizations using artificial intelligence (AI) and machine learning (ML) to support decision-making.
The study, "Data Bias: The Hidden Risk of AI," conducted by independent research firm Insight Avenue, involved interviews with over 640 business and IT professionals, director level and above, from organizations that use data to make decisions and plan to use or are already using AI and ML for decision-making.
Biases inherited from cultural and personal experiences often lead to unexpected and potentially harmful outcomes in AI/ML models. When data is collected and used in training these models, they inherit the biases of those who built them.
Despite the potential legal and financial risks associated with data bias, many organizations need more understanding and resources to address data bias effectively. The lack of understanding stems from training, processes, and technology issues.
The study found that while 74% of Singaporean business and IT decision makers believe data bias will become a more significant concern with increasing AI/ML use, only 16% are currently addressing it through an ongoing evaluation process. The most significant barriers identified include poor awareness of potential biases, a clear understanding of identifying biases, and the need for more available expert resources, such as access to data scientists.
Other key findings from the survey include:
-
58% of organizations expect to rely more on AI/ML decision-making in the coming years.
-
68% believe there is currently data bias in their organization.
-
74% think they need to do more to address data bias.
-
48% view a lack of awareness and understanding of biases as a barrier to addressing data bias.
John Yang, vice president for Asia Pacific and Japan at Progress, emphasized the importance of managing data bias to maintain customer trust and avoid missed financial opportunities and security risks. He noted that one in two Singaporean businesses cite eroding customer trust as the most significant implication of unchecked data bias. Additionally, local organizations are concerned about lost financial opportunities (48%) and security and governance risks (38%).
Progress aims to help businesses make insightful decisions by putting customers at the center of their operations and exploring how AI/ML can enhance effective decision-making that drives businesses forward. By understanding and addressing data bias, organizations can improve the accuracy and reliability of AI/ML-driven decisions, ultimately leading to better outcomes and performance.