By Jason Corso, Stevens Institute of Technology

Work and home life is becoming increasingly dependent on artificial intelligence, or AI. Never before have digital tools been so responsive to us, nor we to our tools. As artificial intelligence continues its transformative impact on every aspect of daily life, it is important to study public attitudes toward this technology to enable its adoption, to ensure it is being applied safely and ethically, and to better understand misperceptions and gaps in knowledge.

To that end, Stevens Institute of Technology conducted the Stevens TechPulse Report: A Perspective on Americans’ Attitudes Toward Artificial Intelligence, a national poll of 2,200 adults led by Morning Consult that examined Americans’ perceptions on a wide range of AI-related issues. The survey found that when it comes to AI, Americans express deeply conflicted views on AI’s use and applications.

More Americans than not believe the positives of AI outweigh the negatives (48% vs. 29%), such as credit card fraud monitoring as well as the advantages of AI being able to handle repetitive tasks and dangerous jobs. But Americans are wary about giving up too much control to AI, especially when it comes to applications that could endanger lives, such as AI in autonomous driving and in healthcare.

In the case of autonomous vehicles, Americans are not sold on the idea. Four in 10 Americans think that AI should not play any role in operating vehicles; only 13% want AI to assist humans and a mere 4% think AI should operate vehicles alone. This is a surprising result. Americans clearly support lane following, parking assist, and obstacle detection capabilities for autonomous vehicles. But the survey results suggest otherwise, reinforcing the need to better educate the public about AI’s current capabilities and uses, and its limitations.

While AI will radically alter how work gets done and who does it, the survey data suggest that the technology’s larger impact will be in complementing and augmenting human capabilities, not replacing them. Respondents were most comfortable with AI playing a supporting role in helping accomplish a variety of tasks, but with humans remaining in the driver’s seat and providing intense value alongside the AI-based technology. There is a long way to go before we truly understand how the symbiotic relationship between AI and humanity will evolve, but the pastures are verdant and a lot of good discoveries to be had.

Top Concern: Loss of Privacy

In return for reaping the benefits of AI, a large portion of the respondents expressed concern about needing to give away certain privileges. The top concern expressed about the technology was that AI would contribute to a loss of privacy. Seventy-four percent of respondents believe AI will lead to a loss of personal privacy, 71% believe it will reduce employment opportunities, and 60% believe it will increase political polarization. Not surprisingly, a majority believe AI will be misused by governments, individuals, and businesses.

However, age affects those opinions: among generations, GenZers are least concerned (62%), while Baby Boomers are most concerned (80%).

But even when it comes to privacy, there is conflicting data.

The use of facial recognition technology, which is highly controversial because of privacy issues, is seen by a strong majority as a responsible use of AI, particularly when it comes to assisting in criminal investigation and finding lost children. In addition, almost half of those polled are comfortable with using it for monitoring people at rallies and marches.

Given people’s concerns about AI, it would seem that they would be inclined toward having the technology better regulated. Our data show otherwise. Americans are evenly split on whether AI is well regulated or not, when in fact, there is actually limited to no regulation by the government right now. These contradictions may reflect the public’s lack of understanding and knowledge of AI and its oversight, and speaks to the need to educate Americans about AI, what it can do and what it can’t.

Artificial Intelligence, despite its promise, remains in its infancy from a capability’s standpoint. For example, take a straightforward task like making a peanut butter and jelly sandwich, which arguably even young children can perform. One imagines that when preparing to make the sandwich, you have the ingredients in front of you along with the utensils needed, and a good understanding of the necessary step. Naturally, one expects an AI-enabled system, such as a kitchen robot, should be able to execute such simple steps like spread the peanut butter. It can’t. It’s so hard. And this is an easy, well-known task. What happens when you want to add banana to the sandwich? Or you ask the AI-system to make a recipe it doesn’t know yet? It would have to watch a YouTube video like a human. Amazingly difficult.

The results of Stevens TechPulse Report: A Perspective on Americans’ Attitudes Toward Artificial Intelligence reinforces one of the main pillars of the Stevens Institute of Artificial Intelligence: to increase AI literacy among students, professionals and the public. Our summary finding about education and broad AI literacy is that we can do a lot more. There is a significant need for education, starting with young children in K-12, as well as in our professional careers across all industries.

AI is a complicated landscape of technology and it’s understandable that while utilizing the technology on a daily basis, most people have very limited knowledge of how AI is currently used and its impact on their lives now and in the future. What is clear is that much more needs to be done to better educate the public about AI, as well as those studying and working in the field.

Jason Corso is the director of Stevens Institute of Artificial Intelligence and the Viola Ward Brinning and Elbert Calhou Brinning Endowed Chair at Stevens Institute of Technology. As the director of SIAI, Corso oversees the long-term strategy and day-to-day operations of the institute as it takes on the challenges of the next five-year plan. With an emphasis on thought-leadership, engagement, and entrepreneurship, he envisions an institute that emphasizes the social relevance and impact of AI in both local and international communities. Corso is also the co-founder and chief executive of Voxel51, an AI software company that helps machine learning and computer vision to rapidly curate and experiment with their datasets in order to build higher performing machine learning systems.