'Democratization' of AI beginning in insurance

Register now

Artificial intelligence, or AI, is seen as the domain of only the most technically trained. However, work is underway to make it more of a tool for business-side managers.

AI has become a big deal in the insurance industry. A recent survey by TCS finds the insurance industry outspent the other twelve verticals surveyed, investing on average $124 million in AI systems, compared to a cross-industry average of $70 million. However, at the same time, Gartner estimates that only five percent of global insurers have invested in or deployed AI at this point.

Overall, companies across all industries in the TCS study reported that their AI initiatives had a strong positive impact on both revenue improvement and cost reduction in the specific business areas where they invested in artificial intelligence. The average revenue increase across all 13 industries was 17%, while the average cost reduction was 12%.

An article in MIT Technology Review looked at efforts to “democratize” AI, suggesting that the day is coming when most managers could learn to deploy AI-based solutions relatively easily.

The article cites the work of Scot Barton of Farmers Insurance, who is charged with tracking customer behavior and policy design. Barton did not have to “hire an army of AI wizards” to do his team’s work with “cutting-edge machine-learning techniques, from deep neural networks to decision trees,” the article’s author, Will Knight, states. Instead, Barton’s team “uses a platform called DataRobot, which automates a lot of difficult work involved in applying such techniques.” DataRobot is at the forefront of a movement – also joined by companies such as Google – to bring AI power to the masses.

The goal of such solutions is to turn AI technology “on itself, using machine learning to automate the trickier aspects of developing AI algorithms,” Knight explains, adding that “some experts are even building the equivalent of AI-powered operating systems designed to make applications of the technology as accessible as Microsoft Excel is today.”

Along with ease of use, these efforts to make AI more accessible also hope to alleviate the growing shortages of data and computer scientists. “A growing number of research papers are popping up on using its techniques to automate more and more aspects of AI.”

While Knight concludes that it’s going to take time until AI is something that someone without training can easily deploy, that day is coming, with the best minds in the industry focusing on automating the AI deployment process.

For reprint and licensing requests for this article, click here.
Artificial intelligence Machine learning RPA