The Promise and Peril of AI and Innovation in Insurance

The integration of AI and machine learning has accelerated exponentially across the insurance industry.

Artificial intelligence (AI) and machine learning tools have been gradually making inroads in insurance business applications over the past decade. However, the dramatic unveiling of ChatGPT in late 2022, along with other easily accessible large language models, has accelerated adoption timelines exponentially across the industry. 

In a recent roundtable meeting of the Datos Insights Insurance Special Interest Group on AI and Innovation, we explored the implications of this rapidly unfolding transformation. Attendees represented a diverse cross-section of life/annuity/benefits and property/casualty insurers eager to compare perspectives and action plans for leveraging leading-edge AI capabilities while mitigating risks. 

The Promise: Massive Productivity Gains 

Early adopter feedback remains overwhelmingly positive. Several participants highlighted the value being generated by AI applications. Use cases that streamline traditionally labor-intensive processes received particularly enthusiastic endorsement: tasks like reviewing new business submissions, extracting details from third-party reports, and summarizing claim adjuster notes, to name a few examples. 

“It can really take [a task that would require] hours and deliver condensed summaries in seconds or minutes instead. Our adjusters can brush up on key events in a fraction of the time before contacting policyholders now,” noted one executive. Their peers concurred, and some insurers shared that they are already seeing productivity improvements as a result. 

The Peril: Risk, Bias, and Regulatory Compliance 

However, AI is not a magic cure-all; insurers must approach it with caution. These tools, while promising, have hardly reached full maturity yet. Consideration must be given to regulations as well as risks like chatbots making unauthorized coverage recommendations. Existing biases in algorithms and training data need to be proactively addressed as well.  

Carrier leadership teams should maintain a balanced perspective. While still in the early days, most seem to believe productivity and customer experience gains currently outweigh the risks posed by thoughtful AI adoption. Yet an undercurrent of healthy caution pervades. As one attendee noted, “This is all moving incredibly fast, and we must be responsive while avoiding recklessness.” 

The Prognosis: Strong Leadership and Enterprise Governance 

Getting governance right may not be as glamorous as deploying AI capabilities, but it’s what separates winners from losers. Governance and risk management practices must evolve to keep pace with rapid application advances already underway within insurance organizations. Moreover, those processes require ongoing oversight rather than once-and-done enactment, given the speed of change industrywide. 

Carriers should incorporate AI strategically into their existing risk, compliance, and IT checks and balances while staying flexible since environmental dynamics keep shifting significantly. The ability to audit model transparency and explainability will be just as crucial as thoroughly scrubbing training datasets. 

Harnessing AI’s immense potential for insurance while steering clear of the pitfalls necessitates asking tough questions together as an industry. We remain committed to helping insurance leaders find the right answers to unleash innovation responsibly. If you would like to join one of our upcoming discussions or learn more about our research, please visit our website or reach out to me at [email protected].