After sustaining elevated claims expense in 2022, the U.S. Property & Casualty Insurance Industry (P&C) is now dealing with the greatest crisis in its long history. Catastrophic weather-related property and casualty losses, driven higher by inflation, are totaling in the billions. Those affected include hundreds of reinsurance and primary insurance carriers; risk managers, independent agents and brokers; and most important of all, millions of P&C insurance buyers!
Company reinsurance is rapidly drying up and retail insurance capacity is vanishing at an alarming rate. Insurance companies are withdrawing from high risk states while limiting new writings. The average home or business owner is finding it difficult to purchase property insurance in the private market at any cost. In the effort to resolve capacity problems and remain solvent, insurance companies are finding long accepted remedial actions are now ineffective. There is a limit to premium increases after all. If this continues on for an extended period, the outcome could be quite sobering for all; perhaps leading to federal political interference and regulation. That alone could destroy our free enterprise insurance system; a superior spread loss mechanism that allows the transfer of a financially unbearable risk of loss to a financially sound insurance carrier at a much lower cost, the premium.
With few exceptions, Insurance observers agree the industry has governed its fiduciary responsibilities rather well over the years. However, for reasons stated above, many insureds do not believe this to be true today. The inability to obtain coverage at a reasonable cost is both frustrating and frightening. The urgent need to return the industry to normalcy is pushing companies to adopt promising, but yet unproven technology; the darling of which is generative (self-learning) artificial intelligence (AI).
Alarms are sounding for the use of AI without sufficient human control. The acknowledged godfather of AI, Geoffrey Hinton, is now warning us of its hidden dangers. Yet, AI continues to be heralded as the golden path to enlightenment. All of this is creating what could be accurately termed a shared human/machine conflict.
Business managers seeking a competitive edge extol its virtues before its abilities are fully known. AI is operating in countless forms in all phases of post modern commerce. This is spurred on by Wall Street investors seeking Apple-like returns among the 57,933 companies labeling themselves AI-involved. So just what is AI doing or not doing to create problems for the industry and its insureds?
AI is now quite capable of detecting and overcoming what it alone judges to be an insufficient or outmoded business model and make its own generative changes. This is creating large numbers of insurance policy processing errors due to dynamic data base incompatibilities it cannot resolve. Without alignment to human values and control, it is also authoring countless pages of unmanageable, robotic-like underwriting guidelines; misused in new pre-quote, pre-renewal and re-underwriting of long existing insurance accounts. Loss of productivity, loss of income, costly processing errors, and yes, higher insurance premiums are the result. This is a prime example of unnecessary shared failure.
To learn on the job, generative AI requires an insatiable amount of information. It must search for, locate, assimilate and process it all. Ironically, this became quite apparent when it was given the task of reducing insurance fraud. Though a worthy goal, fraud reduction allowed AI free reign to gobble up every scrap of information it can find on us, and do so without regard for individual rights of ownership or privacy. In a comedic sense, it is sad AI cannot view itself in the mirror and see the shameless error prone character staring back at it. Because human IT managers advocated its use, they share this failure up front.
The next few months will be critical for the industry. Though we earthlings can do little to control short term weather cycles, we may yet be able to cure AI’s inhuman robotic tendencies. By placing cautious individuals in charge of monitoring and safety management, its inventor Hinton’s hope that AI won’t outsmart us before we can get it under control.