Synthetic data to power 40% of insurance AI by 2027

Addressing privacy and bias concerns through innovative data use.

By 2027, 40% of AI applications are projected to utilise synthetic data to enhance privacy protections and reduce biases in automated processes and decision-making, according to industry experts.

Surya Narayan Saha, Research Manager of Financial Insights at IDC Asia/Pacific, said, "Synthetic data mimics real-world data without compromising individual privacy, which is paramount in a data-sensitive sector like insurance.” 

The insurance industry has traditionally relied heavily on historical data to evaluate claims and assess risks. However, this data can carry inherent biases that may skew AI-driven decisions. 

"Insurers can use synthetic data to simulate diverse demographic profiles and scenarios, ensuring that their risk assessments do not disproportionately impact any group," Saha noted. This method allows for more equitable AI applications, particularly in critical areas like claims management, where synthetic data helps generate varied scenarios to train AI models.

Moreover, the move towards synthetic data is also seen as a compliance measure with evolving global regulations, such as the EU AI Act, which mandates stringent privacy and fairness standards in AI applications. 

Aside from privacy and fairness, synthetic data also plays a crucial role in enhancing AI's capabilities in sales and customer engagement within the insurance sector. By deeply analysing customer data, AI tools can deliver personalised engagement strategies, significantly improving conversion rates and customer satisfaction. 

"AI tools enable insurers to understand individual customer behaviours and preferences, allowing for tailored communications that resonate more effectively," Saha said.

Despite the advantages, integrating AI into existing insurance systems presents significant challenges. Many insurers operate on legacy systems that are not readily compatible with the latest AI technologies, leading to technical hurdles like data compatibility issues and the need for extensive system upgrades. 

"Ensuring the accuracy and reliability of AI-driven insights is another major challenge," Saha pointed out. This requires a robust data management practice to maintain the integrity and completeness of the data used.

To overcome these challenges, Saha said insurers must adopt a strategic approach, investing in technology infrastructure and preparing to embrace technological changes. Continuous monitoring and auditing of AI systems are also crucial to identify and mitigate any emerging biases and to optimise AI performance over time.