FutureAiSummit'24: AI and the future of insurance

Summary

On the Purple Stage panel I moderated at FutureAiSummit'24, we discussed the transformation of Turkey's insurance sector and the effect of AI-based products on the ecosystem. I introduced myself as a board member of the Insurance and Reinsurance Brokers Association and chair of its Digital Insurance Committee, a partner in Genç Sigorta which I have run for fifteen years, and founder of Gensure Tech, which has done AI research and development for two years. On the panel, Prof. Dr. Selim Yazıcı represented the academic angle, robo-advisor entrepreneur Onur the technical and critical one, and lawyer Burcu Tümer the legal one. I set the frame in one sentence: insurance's classic approach was built on detect and repair; with AI we are turning it into a predict-and-prevent strategy.

In my own contribution I said this transformation will affect every stakeholder at its foundation, from insurers to intermediaries, financial advisors and suppliers. Through an intermediary's eyes, I explained that brokers and agents can serve more customers with higher-quality service through AI-based assistants, and that underwriting, pricing, risk profiling and real-time pricing will be reshaped by data. Prof. Yazıcı said insurers hold plenty of tools to turn data into value: hyper-personalisation, dynamic models in place of GLM, IoT-triggered parametric products, region-level life products, and person- and product-level fraud detection on the claims side. He listed analytics' four questions: what happened, why, what will happen, and the real one, what can we do to prevent it.

Onur deliberately opened the opposite window: we are at the very beginning of AI and it is overhyped; the difference between AI-assisted and AI-based matters; ChatGPT began to be written in 2017 and its verbal side is strong while its analytical side is weak; the biggest obstacle is not technology but human resistance; and clean, reliable data comes before everything. Burcu Tümer explained the risk-based approach of the EU AI Act: most insurance applications fall into the high-risk class, which means standardised data sets, user information and complaint mechanisms where a customer can ask why they were scored that way. With the Data Act coming into force in January 2025 we discussed data markets and standard contracts, smart contracts, regtech and a new kind of insurance against wrong decisions caused by hallucination.

In the second half of the panel I turned to risk. According to the World Economic Forum's 2024 report, the most severe global risk of the next two years is disinformation and misinformation, and it sits in the top five of the ten-year list. Generative AI has exploded content production and spread; close to three billion people went to the polls in 2024, and AI-based fraud can damage election results or legitimacy. The question of who is a machine and who is human, and the possibility of states that hold the technology isolating others, were on the table too. On the macro side Prof. Yazıcı noted that flood damage may exceed earthquake in lira terms among catastrophic risks, described detection and prevention through satellite imagery, and a good-farmer insurance prototype running on blockchain, IoT and smart contracts. In closing I told young people: insurance will become a sector that works with data and pushes the limits of AI; for anyone planning a career intertwined with technology, the next five years are a major opportunity.

In this talk

• From detect and repair to predict and prevent: insurance's shift from reactive to proactive
• AI affecting every stakeholder: insurers, intermediaries, advisors, suppliers
• AI assistants for intermediaries: more customers, higher-quality service; underwriting and real-time pricing reshaped by data
• Selim Yazıcı's toolbox: hyper-personalisation, dynamic models, IoT-triggered parametric products, fraud detection
• Analytics' four questions: what happened, why, what will happen, what can we do to prevent it
• The critical view: AI is overhyped, AI-assisted versus AI-based, human resistance and the clean-data condition
• The EU AI Act: insurance applications in the high-risk class; standard data sets, information duties, explainable complaint mechanisms
• The Data Act, smart contracts, regtech and new insurance products against AI decisions
• WEF 2024: disinformation the most severe global risk; close to three billion voters in 2024 and AI-driven fraud risk
• Macro benefit: satellite-based detection and prevention for catastrophic risks, a good-farmer insurance prototype, sandboxes and innovation hubs
• To young people: the next five years are a major career opportunity at the intersection of insurance, data and AI

Insurance's classic approach was built on detect and repair; with AI we are turning it into a predict-and-prevent strategy.