Data Shows How Bias and Lack of Transparency Around Generative AI Impacts CX

Customers are expressing concerns and dissatisfaction about bias in AI algorithms and a perceived lack of transparency in the way generative AI is being utilized, revealed a recent survey conducted by TELUS International.

“It is crucial that companies proactively address biased data and reckless algorithms from the start to avoid severe consequences and inaccurate outcomes. Model validation and tuning are essential for improving the performance and reliability of AI models as they help identify and address potential errors, improve accuracy and ensure that the model can effectively adapt to and make accurate predictions on new, previously unseen data. Additionally, by implementing appropriate policy guardrails, companies can protect customer data and promote a safer user experience while mitigating hallucinations and bias,” said Siobhan Hanna, Managing Director, AI Data Solutions, TELUS International.

Bias prevents AI from giving recommendations and driving opportunities 

As many as 43% of respondents reported that they believe the presence of bias in AI algorithms resulted in them receiving undesirable content, such as music or TV programs they didn't enjoy, as well as irrelevant job opportunities.

In addition, around 32% of respondents expressed the belief that bias present in AI algorithms was responsible for them being denied opportunities, such as the approval of a financial application or a job opportunity.

The emergence of generative AI opens up exciting opportunities for brands to enhance customer experiences. However, for it to be effective, it must provide accurate and safe content. Unfortunately, generative AI sometimes produces inaccurate or nonsensical information, referred to as "hallucinations," which can potentially harm customer loyalty that has been built over time.

Consumers prefer brands that leverage AI  

Based on the survey findings, 40% of American consumers lack confidence in companies utilizing generative AI technology on their platforms, perceiving insufficient measures to safeguard users from bias and false information. Moreover, a significant majority of over 75% believe that brands should conduct algorithm audits to address and minimize bias and prejudice before integrating generative AI into their platforms.

The need for human touch  

The survey revealed that respondents highlighted the need of human involvement, with 49% recognizing that an AI algorithm cannot function effectively without human input. In addition, 19% admitted their lack of awareness regarding the fact that humans review AI algorithms.

“Harnessing human intelligence in a manner that reduces bias is key to successful machine learning. Unlike AI, humans have the ability to understand context and tone, which is crucial to ensuring bias is responsibly mitigated. To effectively reduce bias in AI, companies must source trusted and diverse training data sets that incorporate a wide range of views and perspectives. By adopting a ‘human in the loop’ approach, companies can ensure increased accuracy and reduced bias in its AI datasets,” continued Hanna.

Elsewhere, TELUS has partnered with Amazon Web Services (AWS) to develop a new smart living solution leveraging the latest cloud, IoT, ML, and AI technologies.