From film suggestions to routine customer support inquiries, People now depend on synthetic intelligence to tell client decisions, however new analysis from client and societal options agency MITRE on AI traits finds that lower than half (48 p.c) imagine AI is protected and safe, whereas a big majority (78 p.c) are very or considerably involved that AI can be utilized for malicious intent.
The MITRE-Harris Ballot Survey on AI Traits, performed by The Harris Ballot, additionally finds that most individuals specific reservations about AI for high-value purposes reminiscent of autonomous autos, accessing authorities advantages, or healthcare.
“Synthetic intelligence expertise and frameworks may radically increase effectivity and productiveness in lots of fields,” mentioned Douglas Robbins, MITRE vice chairman, engineering and prototyping, in a information launch. “It will possibly allow higher, sooner evaluation of images in fields starting from drugs to nationwide safety. And it will possibly exchange uninteresting, soiled, and harmful jobs. But when the general public doesn’t belief AI, adoption could also be principally restricted to much less essential duties like suggestions on streaming providers or contacting a name middle within the seek for a human. That is why we’re working with authorities and trade on whole-of-nation options to spice up assurance and assist inform regulatory frameworks to boost AI assurance.”
Given the uncertainty round AI, it’s not shocking that 82 p.c of People—and a whopping 91 p.c of tech consultants—assist authorities regulation. Additional, 70 p.c of People—and 92 p.c of tech consultants—agree that there’s a want for trade to make investments extra in AI assurance measures to guard the general public.
“Whereas we see variations by gender, ethnicity, and generations in acceptance of AI for each on a regular basis and consequential makes use of, there stays concern about AI throughout all demographic teams,” mentioned Rob Jekielek, managing director, Harris Ballot, within the launch. “Males, Democrats, youthful generations, and Black/Hispanic People, nonetheless, are extra comfy than their counterparts with using AI for federal authorities advantages processing, on-line physician bots, and autonomous, unmanned rideshare autos.”
Different key findings embody:
- Three-quarters of People are involved about deepfakes and different AI-generated content material.
- Lower than half (49 p.c) can be comfy having an AI-based on-line chat for routine medical questions.
- Solely 49 p.c can be comfy with the federal authorities utilizing AI to help advantages processing.
MITRE is collaborating with companions all through the AI ecosystem to allow accountable pioneering in AI to raised impression society, together with superior modeling capabilities for AI assurance to deal with the sophisticated impact of a promising expertise’s potential impression on techniques and society. The agency participates in a number of joint collaborations, together with membership within the Partnership for AI and Era AI Consortium.
Entry the complete report right here.
This survey was performed on-line inside the US November 3–7, 2022, amongst 2,050 adults (ages 18 and over) by The Harris Ballot by way of its Harris On Demand omnibus product on behalf of MITRE. Tech consultants have been surveyed in October 2022.
