Artificial general intelligence (AGI) is expected to replace today’s fragmented AI ecosystem with a single system capable of handling underwriting, actuarial analysis, and compliance functions, according to Reinsurance Group of America (RGA).
Insurance AI today operates through a network of specialized systems connected through conventional programming. This structure, often described as agentic AI, enables coordination among models designed for distinct tasks, allowing them to execute workflows and interact with external data sources.
Each system performs a narrow function. Models are built separately for text generation, image creation, speech recognition, and code development, while others support translation, autonomous driving, and medical image interpretation. These systems can be combined, but they do not function as a single adaptable model.
This architecture creates an experience that appears integrated but depends on multiple layers of coordination.
AGI refers to systems capable of performing tasks across domains without requiring manual integration of separate models. Unlike current AI, which is designed for specific functions, AGI would apply learned capabilities to new problems without additional programming.
The transition would resemble the consolidation of multiple tools into a single platform. In insurance, this could affect underwriting, actuarial projections, regulatory reporting, and capital modeling, reducing reliance on interconnected systems.
AGI is not expected to exceed human intelligence in its initial stage but would represent a more general capability than current systems. Further advancement toward superintelligence would depend on recursive self-improvement, where AI systems independently refine their own programming.
Read next: Insurance executives fear AI bias - survey
More than three years after the release of ChatGPT on November 30, 2022, AI tools have become widely used across insurance operations. Functions once limited to actuaries, data scientists, and software engineers are now accessible to a broader workforce, including non-technical staff.
This shift has changed how work is performed and who participates in it, with AI supporting tasks such as medical data review, document processing, and communication drafting. While these tools appear unified, they rely on multiple systems working together behind the scenes.
Industry research indicates that adoption is increasing. Surveys show 90% of insurers plan to increase AI investment, with 75% focusing on underwriting and claims functions.
AI is already used to process large datasets, inform pricing, and support claims decisions in near real time. Tasks such as loss analysis, application review, and claims documentation are increasingly handled by AI systems, allowing staff to focus on advisory, client interaction, and oversight.
For reinsurers, these developments extend to portfolio analysis and risk aggregation, where data processing capabilities influence pricing assumptions and capital allocation.
The limits of current AI are evident in task-specific performance. Systems designed for chess can outperform human players but cannot produce written analysis.
Language models capable of writing text are not optimized for gameplay and may perform poorly in that domain. This separation highlights the constraints of systems designed for single tasks.
AGI could alter how risk is assessed and priced by enabling more individualized analysis and shifting attention toward loss prevention. At the same time, operational roles such as underwriting and claims handling could face increased automation, depending on regulatory frameworks.
Regulatory oversight is already evolving. US states, including New York, have introduced requirements addressing transparency and bias in AI systems. Human oversight, accountability, and compliance remain necessary components of AI deployment.
The ability to deploy advanced systems at scale also introduces competitive considerations, where early adoption could influence operational efficiency and decision-making capabilities across the market.
Forecasts for AGI vary. Metaculus Prediction has estimated a median timeline of May 2033, based on a question first posed in 2020. Other projections suggest a 50% probability of development by 2061.
For insurers and reinsurers, AGI represents a potential transition from interconnected systems to a unified model capable of handling multiple functions. Current deployments, including agentic AI, indicate that the industry remains in an intermediate stage.
Ongoing developments in data availability, regulatory frameworks, and system capabilities will influence how quickly that transition occurs. AGI, if achieved, would change how insurance operations are structured, moving from integrated systems toward a single adaptable framework.
