Insurtech funding transactions have seen a rise in the second quarter of 2018 with life and health (L&H) insurtech reaching $5bn. The 71 transactions mark a record high for the sector – due to AI advances making more data available for L&H. The latest briefing, compiled in collaboration with machine intelligence platform - CB Insights - also reveals the expansion of new forms of underwriting data because of advances in the understanding of the human body and the increase in the use of wearable sensors that track activity and monitor behaviours. Narrow AI & GAI Insurance companies need to avoid the broad AI definition that is “hopelessly confounded in its own hype”. AI is divided into two types – General Artificial Intelligence (GAI) that is capable of human level cognition such as self-directed problem solving - and Narrow AI which is task specific and non-recognisable. However, the sector should ignore GAI, as it does not currently exist and there are doubts as to whether it ever will, argues the report. On the other hand, Narrow AI must gain the sector’s focus as it can fall under software or predictive models. “For insurers the applicability of AI is based on narrow AI technologies as natural language processing and image recognition are both machine learning implementation with working business applications such as predictive models,” added Eaton. Unlocking insurance methods Many of the modeling techniques that fall under the AI umbrella are classification algorithms that require data. “AI therefore is the key to unlock the door of big data” for insurers, it suggests. However, “unlocking the power of these methods requires a sufficient volume of training data.” This data can come in a variety of forms and does not need to be insurance centric. Such third party data can include aerial imagery to determine building characteristics or estimate post-catastrophe claims potential. Other data that may be external to the industry is behavioural data built on the interactions of users with digital platforms such as social media and web search. Data directly related to the insurance industry that may be unlocked includes data sets in claims.