Controlling the Future
It can be both incredible and terrifying to witness a technology once only conceivable in science-fiction become increasingly intertwined with modern society. While “AI” has been a buzzword in all professions and in everyday life, it is difficult to bear in mind that this technology is not new, it is just rapidly evolving to higher capabilities, as predicted. AI, or artificial intelligence, has been utilized in GPS tracking and maps for years, as well as in systems like Siri and Alexa.
Even when voice-detecting software that spoke back began to emerge—and for decades prior—many held onto a Hollywood-manufactured idea that humans would be replaced or endangered by innovation that gained sentience. Any new tool can be used for nefarious purposes, and while the nature of AI obtaining its knowledge from the human race can cause skepticism, experts are actively working to repurpose it for the greater good. Furthermore, with the prior experience of the early internet acting as a Wild West for connecting people globally—for better or for worse—we now know that powerful technology requires regulation.
Experts are anticipating how artificial intelligence can provide a better quality of life, knowledge and efficiency. Meanwhile, many are considering the negative potential and current impacts from AI such as falsified media, job replacement and the capacity for bias. In South Jersey, professionals from all different fields are witnessing the growth of AI firsthand, while also providing their knowledge and experience of utilizing AI to provide care and support rather than to do harm.
Health Care
South Jersey is known for its state-of-the-art health care because of its systems’ propensity to study, to utilize and to continuously develop technology. Virtua Health exemplifies incorporating AI in a humanistic way, in that it strives to care for patients both physically and emotionally and works collaboratively with doctors. The capability of AI to complete repetitive tasks allows medical professionals to focus their time and energy on addressing patients’ needs.
“We’ve really focused around the intelligent automation side of artificial intelligence, and more prediction and support tools. We’ve utilized AI to do things like predict length of stay, to predict sepsis potential of a patient so we can get early intervention,” says Tom Gordon, Virtua’s chief information officer. “We’ve also invested time and effort into ambient voice from an artificial intelligence perspective … [implementing] a tool that is in the physician’s office that listens to the conversation, breaks that conversation down, writes the note, and the provider still has to validate that the note was right and sign it.”
Virtua’s Chief Digital Transformation Officer Dr. Tarun Kapoor adds that a trial of reading 6 million progress notes, which would take a human 27 years to complete, resulted in the machine deciphering them in one weekend.
Virtua recognizes Gordon and Dr. Kapoor as key contributors to its AI-related research and projects, and both stress that this innovation will not replace medical professionals, but rather help them expand their focus beyond menial tasks with more efficient processing. In addition to the ambient voice element recording and deciphering conversations, the technology allows for doctors and nurses to loop in patients’ family members, creates appointments and helps patients access a nurse faster when every second counts.
The South Jersey health-care system also incorporates a chatbot (through a partnership with Woebot Health) that helps people with anxiety or depression self-administer cognitive behavioral therapy, expanding technology that was once easily identifiable as a robot with its disjointed and awkward speech into something that provides meaningful, human-centric or emotionally resonant advice on demand—and reaches those who may not have access to therapy. Dr. Kapoor points out that peak usage is between 2 and 5 a.m., and Gordon affirms that patients’ needs don’t cease to exist after business hours. This virtual therapist is intended to be a completely neutral entity that exists exclusively for mental-health purposes, which may be more appealing to those who have anxiety about sharing with a person who may or may not have internal biases.
Of course, as with any brand-new offering, this is not an infallible solution. Dr. Kapoor stresses that machines still make mistakes, and Gordon likens it to the frequency of GPS systems providing inconvenient routes instead of the obviously faster ones. Still, the exponential development provides hope for the future of effective treatments and cures previously inconceivable.
“Computers can help us understand and decipher things. I think based off that, it’s going to lead to therapies,” says Dr. Kapoor. “I’ll just throw a number out there that is pretty mind boggling: In the beginning of the 20th century, medical knowledge was doubling every 50 years. And so a physician would learn a brand-new set of knowledge over 50 years. Medical knowledge is now doubling every 45 days.”
Education
In the same scope of pioneering life-saving solutions that previously met roadblocks from human limitations, AI software is becoming an integral part of school safety. Rancocas Valley Regional High School in Mount Holly became the testing ground in 2018 for the AI firearm-detection software ZeroEyes, an initiative spearheaded by Superintendent Dr. Christopher Heilig.
ZeroEyes is a multi-layered security system trained to recognize firearms on camera, and alerts the ZeroEyes team so they can verify it and rapidly contact law enforcement. It is developed to identify firearms in different positions, locations and even in a pitch-black room.
Unfortunately, the mass gun violence epidemic in America has been met with zero solutions to halt the deaths of schoolchildren. Since the massacre at Columbine more than 20 years ago, drills have been regularly implemented in schools, but taught students to hide under tables, or more recently, to fight back when confronted by a gun. There has been widespread criticism for decades over the incident response times, as well. With ZeroEyes, active-shooter drills have demonstrated outstanding results that cut the response time in half at a minimum.
Sam Alaimo, CRO and co-founder of ZeroEyes, works with a group of fellow Navy SEALs to proactively combat gun violence following the 2018 Marjory Stoneman Douglas shooting in Parkland, Florida. Once again, an essential part of this AI technology is human input.
“From the moment a gun is seen in front of the camera to the moment Dr. Heilig gets a notification on his phone is about three to five seconds. In these situations, obviously seconds matter. From the data we’ve collected, about 68% of mass shootings in K-12 happen outside the building, not inside the building. That’s before they ever even walk up to a metal detector. We want to let you know that there’s an assault rifle in front of an elementary school; that is our sole task and purpose. … We’re plugging into the back end through existing security cameras. We find most schools already have the right cameras in the right locations, so we’re literally just a software component.”
A much less dire impact of artificial intelligence on students is the ability to use it to both write essays and complete assignments. At Rancocas Valley, though, this technology is blocked on campus, and only propels school administrators and staff to recognize and to learn from the adapting educational experience.
“You don’t want AI to do your work, that’s not learning; however, there’s probably a way—and I think every organization across the country is trying to figure this out—to use it to work smarter,” says Dr. Heilig. “We have one of the experts in the country for something called differentiated instruction, Peter Grande [executive director of TLS, or Teaching Learning Succeeding], working with our staff for instructional tools in the classroom, instructional strategies. And they’re talking about how, as a teacher, I can use AI to come up with examples and strategies for my students.”
Entertainment & Social Media
Similar to the evolution of speech imitation and expansion that AI has conquered in such a short amount of time, AI-generated images that used to be easily identifiable as fake are becoming more realistic. Not too long ago, anyone on the internet could type in a prompt to generate an AI image through certain apps or sites, the images produced would seem unnatural, such as their proclivity to design extra fingers, rows of teeth or warped backgrounds. Now, though certain attributes can be spotted as fake by the person in the image, they look real enough to convince the general public otherwise.
In 2019, convincing yet unsettling videos circulated of a deepfake of comedic actor Bill Hader on a talk show, in which he morphs into other celebrities during his famous impressions. While this does not necessarily encompass harmful content, it demonstrates the capacity to lose an audience’s trust with these realistic imitations, something that could be pivotal when expanded to government proceedings and elections.
Another such example is South Park creators Trey Parker and Matt Stone investing $20 million in an AI entertainment startup in 2022 after releasing a comedic web series with a deepfake of the former president with a different hairstyle. As the pair is known to be on the cutting-edge of all types of media, it proves that AI has a potential to expand art and entertainment beyond previous limitations. Still, this demonstrates how easily a person’s image could be changed and utilized maliciously for something that is not harmless, funny media that is easily identified as a replication.
A major concern is the reprehensible act of manipulating innocent images into explicit content. Many women have come forward on social media with nightmarish experiences of finding their bodies recreated by AI with their faces recognizable, causing distress about the impact these pictures could have on their lives if discovered by someone they know. It is an issue for private citizens, especially minors, on social media, and of course this risk expands to news, politics and public figures.
Earlier this month, explicit pictures of Taylor Swift generated by AI were circulating on social media, igniting the demand for stronger restrictions through legislation. Although there already are some safeguards in place, users are able to bypass them and share their techniques on niche, infamous online communities.
Legislation
The legislation surrounding artificial intelligence is still in development, and the potential issues that need to be addressed require a multifaceted approach. There is likely to be opposition regarding copyright infringement or defamation from using a person’s image or brand to create a false representation. However, the comedic examples listed above are likely protected under parody law and categorization of a public figure as the subject, and any laws would strive to discourage purposefully misleading material.
State Senator Troy Singleton introduced a bill to ensure AI technology isn’t abused, and has an implemented checks-and-balances system that appoints a committee and guardrails—a sentiment echoed by Gov. Phil Murphy, who established an AI Task Force October 2023.
“The current bill that we have, which is Senate Bill 1438, is in a committee hearing right now,” says Singleton. “It’s my hope, as we are now in the beginning of this legislative cycle, that we’ll be able to see that bill move and keep the discussion going around the benefits of AI, and recognize that sometimes these technologies work opaquely; their impact on government can be unknown.”
As conversations provoke a drive to action about these topics, AI’s potential to cause more good than harm is at the forefront. It is evident that it could both save lives and negatively impact them simultaneously, depending on who is using it and what their intentions are. That is why it ultimately requires human surveillance at all times, as every expert that South Jersey Magazine spoke with emphasizes.
“I think we should all use AI as informed consumers, who learn how the systems are built and understand the principles of companies building them; I think that is important,” continues Singleton. “And I think if we continue to train developers on AI systems, on diverse and more representative datasets that use fair and unbiased algorithms, then I think that ‘myth’ as it were, can be dissipated around the biases of the AI usage. A lot of that is going to take all of us to make sure, like in human nature, to understand that some of us—all of us—come with some level of preconceived biases that we have to push aside in order to go to a deeper level of understanding.”
Click here to subscribe to the free digital editions of South Jersey Magazine
To read the digital edition of South Jersey Magazine, click here.
Article continues below
advertisement
Published and copyrighted in South Jersey Magazine, Volume 20, Issue 11 (February 2023)
advertisement
Published and copyrighted in South Jersey Magazine, Volume 20, Issue 11 (February 2023)
For more info on South Jersey Magazine, click here.
To subscribe to South Jersey Magazine, click here.
To advertise in South Jersey Magazine, click here
To subscribe to South Jersey Magazine, click here.
To advertise in South Jersey Magazine, click here
Author: Carly Murray
Archives
Who’s Who in Health Care
Ahead of the Curve
Money Matters
Going Greener and Cleaner
Self-Made Man
Building Toward the Future
On The Move
Firing Up
Drinking it All In
Caring for One Another
What’s the Big Idea?
Making Connections
Up to the Challenge
Today’s Pupils, Tomorrow’s Professionals
Best of Home & Garden 2024
More...
