Artificial intelligence (AI), machine learning, and big data have already disrupted conventional health tech research and development in significant ways; many of our posts on Health Tech Insider mention the technology, the algorithms, and the bits and bytes. “Big data” isn’t a new concept, but accessible big data and technology that is sufficiently powerful, agile, and meaningful open a big box for researchers, developers, and manufacturers. Whether history will judge massively accessible data a treasure chest or a modern Pandora’s Box remains to be determined.

Experts from the Pew Research Center and Elon University‘s Imagining the Internet Center explored the question of AI’s net value. During the summer of 2018, the group asked “979 technology pioneers, innovators, developers, business and policy leaders, researchers and activists” for their opinions. The basic question was whether the respondents think that “advancing AI and related technology systems” will “enhance human capacities and empower them” or “lessen human autonomy and agency” by the year 2030.

Overall, 63% of the thought leaders “are hopeful” AI’s effect will be positive, and the remaining 37% said the result would be a net loss or at most no gain. The Pew Research report is a hefty read, with quotes and call-outs from many of the respondents. It cites some of the major concerns posed by AI, such as the “black box” problem in which self-learning systems may not reveal how they reached their conclusions. The tools can be used to help people or take advantage of them. Some predict that AI could result in the loss of a significant number of jobs by automating knowledge work much the way that the Industrial Revolution eliminated manual labor jobs.

On the other hand, some respondents see solutions to the risks of AI, such as creating policies that put a priority on humans over machines. The systems could create new efficiencies that could open up opportunities for new jobs that are more engaging and rewarding. And applied to healthcare, AI could reduce errors and uncover more effective treatments.

We also need to consider that the next step from machine learning, neural networking, is barely past the neonatal stage. When machine learning moves beyond prediction based on correlation to the point where AI determines direction, it will be interesting to see what the experts think at that point.