For professionals, managers or sales and marketers, companies buy AI to enhance humans

The headlines still talk about AI job Armageddon; widespread job loss through AI and automation because AI is faster and better at everything humans can do. But emerging evidence tempers concerns about machines replacing professional work.

Commercial AI applications offered by startups are being bought for enhancing human capabilities rather than replacing humans, according to new research published by Boston University School of Law. Customers are using AI to create jobs in certain occupations about as often as they are used to eliminate jobs. AI applications are more often increasing employment in managerial, professional and sales and marketing roles but are reducing employment in service, clerical and manual roles.

Companies buy AI from startups for a host of reasons. Making better decisions tops the list, with data management and product and service innovation following closely behind.

Benefits to customers of AI products bought from startups

According to the researchers: “Of the survey responses, 54% strongly agree that their products automate routine tasks and 50% strongly agree that their products reduce labor costs. Taking these categories jointly, 55% of firms strongly agree that their products benefits customers by replacing labor. In contrast, counting the remaining benefits as enhancing human capabilities, excluding “Reduce other product costs”, 98% of firms strongly agree that their products benefit customers by enhancing capabilities. That is, AI appears to be much more about enhancing human capabilities than about replacing them.”

The type of AI matters. Many more firms develop their own software for the most commonly used technologies rather than purchase them from an external vendor. Only in two areas do firms rely more on outside vendors: speech recognition with 19% using external products while 13% develop their own, and natural language translation, with 17% using external software and 13% develop their own. By far the most common algorithms are deep learning based, with 76% of reporting firms use neural networks including recurrent, convolutional, and generative adversarial neural networks. 

It’s becoming increasingly clear that it’s the combination of learning machines and humans that will lead to productivity improvements. This research highlights how the startup ecosystem and product sales reinforces this but also that the impact is uneven across industries and jobs.

Share on email
Share on facebook
Share on linkedin
Share on twitter