Doing AI the right way
As the era of artificial intelligence dawns, Cal Al-Dhubaib is ready to help us meet the startling new normal. We’re going to need him.
By Robert L. Smith
Photos by Roadell Hickman
People awoke this year to news that artificial intelligence systems are way more powerful, and omnipresent, than any of us had imagined. Suddenly, a computer can write a research paper, crunch code, draw a painting, even score well on the Bar exam — all with a skill that seems human. When did this happen?
The collective astonishment is understandable. AI has been a brewing force for years, but its sudden utility — and seeming superpower — surprised even Cal Al-Dhubaib ’16, the Case engineer who helped pioneer the science.
Al-Dhubaib (pronounced “Al DU-babe”) earned the first data science degree conferred by the Case School of Engineering in 2016, then went to work at the field’s frontier. His company, Pandata, helps businesses and institutions use machine learning tools to put their data to work solving problems and finding new efficiencies. That’s an enticing prospect to an array of major employers. The Pandata portfolio includes University Hospitals, advanced manufacturer Parker Hannifin, tech firms like Hyland Software, utility giant FirstEnergy, and nonprofits like the Cleveland Museum of Art.
Al-Dhubaib knows as much as anyone about what artificial intelligence can do today and what it might be able to do tomorrow. So when he tells you the latest burst of AI innovation leaves him “terrified and hopeful,” you feel a little terrified. And a little hopeful. He has that kind of effect.
Slight and soft-spoken, the 31-year-old entrepreneur speaks with a pleasant calmness, often at length, as he puts technical complexities into common English.
“None of the math is new,” he says matter of factly. “That’s the thing most people don’t realize. What’s new is our ability to provide enough data for these very mathematical-hungry models to be able to learn.”
He’s saying this inside a co-working office space in Cleveland’s trendy Ohio City neighborhood, in a high-ceilinged room furnished with tables and couches. There’s a juice bar/cafe and big windows that let the sun stream in off Jay Avenue. Al-Dhubaib, like his employees scattered around the country, works remotely, tapping the Wi-Fi wherever he’s conducting business.
Today, the industry focus is on “large language models,” or LLMs. The best know is ChatGPT, a so-called generative chatbot that scrapes the Internet for content it uses to answer questions. The tool startled even computer scientists when it was released in November by a San Francisco software firm called OpenAI. Suddenly, students everywhere had an app that could write their book report — at some risk. The chatbots are known to make mistakes, spread disinformation, and reflect prejudices and stereotypes that lurk in the Internet.
ChatGPT was soon eclipsed by GPT-4, a more powerful version that reportedly scored in the 90th percentile of test takers on the Uniform Bar Examination. Microsoft and Google are competing with similar products and it’s safe to say, Al-Dhubaib concedes, that the world as we know it will never be the same. Still, he’s only modestly impressed with the chatbots.
“It’s nothing more than autocomplete on steroids,” he says. “It’s been exposed to 40 billion documents.”
AI is not smarter than us, he says reassuringly, but it can fool us into thinking it is.
“What’s cool is that it’s really good at these language-based tasks,” he continues, eyes shining. “And it can really accelerate a lot of things” — like writing, drawing, programming. “And it’s only going to get better.”
So are we overreacting to AI’s power, its ability to change our jobs, our lives?
“Yes and no,” Al-Dhubaib says. “What I think the world is going to learn is that there’s a lot it can’t do. And then there’s a lot you don’t want it doing.”
He’s the man of the moment — and not a moment too soon.
Using AI, Cal Al-Dhubaib was able to show the Cleveland Museum of Art how visitors traverse the museum.
Man of Many Worlds
Born and raised in Dhahran, Saudi Arabia — a port city on the Persian Gulf — Al-Dhubaib’s move to Cleveland in 2009 for college was a bit of a homecoming. His mom is from the Cleveland suburb of Maple Heights. She met dad, a retired engineer for the Saudi national oil company, at the University of Akron when both were students. Growing up, Al-Dhubaib spent enough summers with his mom’s family in Streetsboro that “I even have a Cleveland accent.”
Single and adventuresome, he’s a world traveler accustomed to living out of a suitcase. He’s already visited 25 nations toward his of his goal of discovering 50 before he turns 50.
That restlessness was evident during his student days. At Case, Al-Dhubaib jumped from one course of study to another. He had switched majors six times when he told his exasperated parents that he had finally found his vocation, something called data science. Today, it’s part of the Department of Computer and Data Sciences in the Case School of Engineering. In 2016, Al-Dhubaib’s course work became the foundation of a new major.
He’s still thankful to the Case mentors who helped him to pursue his dream, including Professor Peter Thomas and Professor Emeritus Ken Loparo, PhD ’77.
“He had this interest in problem solving and doing it with data,” said Loparo, then chair of the Department of Electrical Engineering and Computer Science. “He had that ability. So we put together a custom program for him.”
A career blossomed. Al-Dhubaib’s first startup, which tried to apply big data to healthcare, failed quickly. Pandata, which he launched while still a student, seized an opportunity presented by the digital age. Companies and institutions were amassing troves of data with little idea of what to do with it. Enter the data scientist.
Putting data to work
The Cleveland Museum of Art offers a good example of the revealing power of big data. Every time you walk through the museum, the Wi-Fi system detects your smart phone, Fitbit, or smartwatch. Every day, hundreds of sensors ping off thousands of personal devices as people traverse 61,000 artworks, leaving an unseen trail like electronic breadcrumbs.
Using machine learning techniques, Al-Dhubaib and his team reconstructed how visitors navigate the museum. They revealed, among other things, the beguiling influence of ARTLENS, the interactive gallery near the museum entrance designed to preview the collection inside. Pandata’s team showed that visitors who spent at least five minutes in ARTLENS spent up to an hour longer in the museum, on average, than other visitors.
Curators were enthralled. Such findings lent insight into visitor motivations at scales not previously possible, said Jane Alexander, chief digital information officer for the museum.
“He showed us pathways, and that is like really exciting, insightful information,” she said. “What I love about Cal is that he’s up for taking on a new challenge and thinking about it in new ways. And he takes a research approach.”
The museum work made headlines, but Pandata does much of its tinkering in what Al-Dhubaib calls “high-risk environments,” like critical care, high finance, and cybersecurity.
University Hospitals offers an example. To better fight cancer, the hospital system asked Pandata to identify patients most likely to be readmitted to its Seidman Cancer Center — and provide insights into what brought them back. After creating a way to anonymously sift patient records, Pandata built a “predictor” for staff to use to identify patients more likely to experience complications.
The new insights are saving money and lives, said David Sylvan, President of UH Ventures, the innovation and commercialization division of the hospital system. He expects to be using Pandata again. “AI is the next frontier for all of us,” he said. “It’s going to be ubiquitous.”
Corralling a runaway industry
Pandata is now six years old and Al-Dhubaib sees 2023 as a big year. He expects revenues to reach about $2 million, up from $1.2 million in 2022. He’s looking to add to his 10-person staff, which is peppered with Case talent.
Niki Agrawal ’16, a Pandata data scientist in New York City, met her boss at commencement. Carlin Jackson ’15, MBA ’16, befriended Al-Dhubaib during their Case days and now consults for him. He recalls his classmate designing his own major before the field of data science really existed.
“It was very cool to see someone do that,” Jackson said. “We all had a general idea of what it is, but we had no idea how integral and important it would be.”
Jackson, the founder and CEO of the business consultancy Theo. Wyes David, Ltd., thinks Al-Dhubaib is still showing that foresight.“He’s just such a savvy communicator of complex topics and ideas,” Jackson said. “When you’re able to take what’s cutting edge and make it common sense, that’s powerful.”
Al-Dhubaib has emerged as a voice for ethical AI, preaching a need for adding guardrails to a mushrooming technology. It’s a theme he has sounded at industry conferences and as a member of the Responsible AI Institute. He thinks data scientists need to work with lawmakers to design rules of the game and quickly.
He’s also aware of how hard that will be. Thanks to the chatbots, AI is approaching mass adoption at astonishing speeds. Some jobs are going to disappear. Unintended consequences loom. The potential to use the technology for dark arts, like weapons, concerns him. He thinks Ohio could take a lead role in shaping regulations and become a hub of human-centered AI innovation and application.
His clients are already asking, should they be adopting this technology in some way?
“Probably,” he says.
He thinks the rest of us need to get into the game, too.
“My argument is, AI is going to evolve as a co-pilot to daily tasks, and accelerate writing, and entertainment, and architectural drawing, and biology. Yes, the scientist still has to run the experiment, but it’s helping you explore that space. It’s not that scary”— unless you’re left out.
As the word processor made the typewriter obsolete, AI is going to change many of the ways we accomplish tasks.
“It should be scary if you don’t adapt your work,” Al-Dhubaib warned, adding softly, “We can do this.”
The interactive ARTLENS Gallery at the Cleveland Museum of Art appears to be as beguiling as curators hoped it would be. Photo by Cleveland Museum of Art.