Challenges around the application of AI, air quality concerns, and the effect of rising inflation on insurance claims, are just some of the issues currently facing the construction sector as it grapples with global uncertainty and change, writes Aon’s Michael Earp.
“2024 is a year marked by rapid technological advancements, global challenges and significant shifts in geopolitics”. This was the response to a recent question to an AI app when it was asked to sum up 2024 in a short sentence. The answer is not incorrect and indeed is indicative of some of the challenges that will face society as more reliance is placed upon AI (machine learning), as the world deals with the hangover of the pandemic. One of Aon’s clients told us earlier this year that of the 144 countries in which they operate, 100 will have elections within 12 months and already we have seen government changes in both the United Kingdom and France.
AI – who’s checking the checker?
Artificial intelligence comes in various grades from the large language model “ask me anything” to smaller specialised systems designed to deliver specific task outcomes. AI is a digital system and it can only review the data that is available to it. It does not have creativity, intuition and common sense. It can draw unexpected relationships between data points that can be insightful and previously unrealised and it may accept and integrate erroneous data.
At Aon we have been thinking about “Generative AI” versus just “AI”. The difference would be that Generative AI has a creative element (it can determine outcomes and draw conclusions), rather than regular AI, which to our knowledge really is simply sourcing and compiling information. AI cannot tell you what the entity’s risk manager should be worried about, just what’s been published and is harvestable from the internet. Obviously, it’s evolving quickly and we’re all still learning about it, but so far AI to us seems mostly to be about interface (asking a central source a question) and it is producing a regurgitation of harvested information previously published by others.
Results from querying artificial intelligence systems should always be verified and this may be where ‘the rubber meets the road’ in the near term. For example, an engineer could query an AI engine for specifications for a certain steel truss – length, dimensional load variation, environment, seismic resistance, etc. At Aon, we believe that output must be verified. We are unsure if querying multiple AI sources for agreement is thorough protection, or that it will necessarily “quickly detect” errors.
Several AI providers could put forth the same incorrect output. In the present nascent state of AI, it should be checked against historical methods, perhaps pen and paper or proven specialised software packages. The engineer is staking their professional licence, reputation and professional liability insurance on their own output. Relying on AI may be a reasonable place to start and indeed could flag matters that the engineer would not otherwise conceive, but relying solely on AI output could be perilous. For now, the maxim might be: “Don’t trust, verify!”
Air quality challenges in a post-Covid world
With new strains of Covid continuing to appear, and flu season just months away, what are the lessons learned about the prevention of the spread of respiratory infections? Viruses like Covid-19 can travel much further than six feet and social distancing, hand washing and masking wasn’t enough. Air quality scientists say that, from the start of the pandemic, there should also have been a focus on improving the air we all breathe indoors. The release of virus particles into the air is going to drift around and not just fall to the ground. Recent focus has been on HVAC systems that might be turned off at night, or if the temperature has reached a certain pre-set level, when there may be very low ventilation.
A recent survey of facility managers in the US and Canada found that since March 2020, roughly two-thirds of respondents have upgraded their MERV (minimum efficiency reporting value) filters and increased their air exchange rates. Engineers need to be aware that standards for air quality may be changed and engineers should know about these changes, for new and existing buildings. Recently, Aon is seeing an uptick in vertical construction and so the air quality of those buildings may be in the forefront of owners’ minds.
Keeping an eye on claims inflation
During the 2022 and 2023 professional liability renewal cycles, there were significant increases in economic inflation. Insurers faced the question, “to what extent would claims costs also inflate?” There is a possibility of claims inflation following economic inflation in most lines of general insurance. For professional indemnity, it takes longer for the consequences of claims cost inflation to become apparent. Costs of defending claims can often be up to 40% of the total claim costs based on our internal Aon analyses of historical claims.
In the current elevated economic inflationary environment, professional service firms have needed to review the possibility of increased premiums and retentions, assess overall limits purchased and review and monitor changes in contractual obligations. It is important therefore for risk managers to review the adequacy of their firm’s total limits in the current and expected future environment. Regularly conducting limit adequacy studies using benchmarking and other data-driven metrics around industry claim data is always recommended.
Increased diligence should be exercised around customer and client contracts and any changes in limits of professional liability being requested by professional service firm clients. There has been a general uptick in contractually required limits and some of this may be due to concerns around inflation.
Michael Earp is the managing director, construction and insurance, at Aon.
