It began as a pandemic-era tool bookmarked by internet browsers, crawling along television news chyrons and cited in local public health updates. More than 2.5 billion views later, the Johns Hopkins University COVID-19 dashboard’s lasting, yet lesser-known, legacy may be how it helped attract monumental funding to build an artificial intelligence hub in Baltimore.
The university’s big bet on artificial intelligence is already beginning to take shape in the city. Over the summer, administrators unveiled an ambitious plan to build a Data Science and AI Institute on the Homewood campus with the goal of positioning Hopkins, and subsequently Baltimore, as the nation’s foremost destination for data science, machine learning and AI.
Officials say the investment will help researchers better harness the university’s deep well of data — and fuel discoveries for the benefit of humanity.
“We’re sitting on gold mines,” said Denis Wirtz, vice provost for research at the university, of the institution’s robust datasets. The challenge now, he said, is making them more user-friendly so researchers across a broad range of disciplines as well as communities can tap into their potential.
The initiative comes at a time of rapid expansion in the field of artificial intelligence and about a year after OpenAI launched its popular language model-based chatbot, ChatGPT. These developments have inspired a new wave of theorizing about AI’s untapped potential — and possible shortcomings.
Hopkins’ academic community has ample experience working in data science and artificial intelligence to identify strokes using facial recognition software and predict Major League Baseball’s most valuable player. They’ve also uncovered shortcomings in the field by collaborating on a study that found racist and sexist stereotypes in popular artificial intelligence algorithms. Some had campaigned internally for years for the university to commit more money to exploring the fields, Wirtz said.
One of the university’s greatest data science victories would come in January 2020, when the Johns Hopkins Center for Systems Science and Engineering launched an interactive dashboard to track the spread of the coronavirus across the globe. The open-source dashboard’s world map culled public health data into a central location, depicting cases and fatalities as little red dots peppering countries and continents.
Soon, the university’s Applied Physics Laboratory joined the project as the scale of data and demand increased. By March, the page attracted more total visits than the websites of the U.S. Centers for Disease Control or The New York Times, the university reported at the time. The dashboard expanded again to become the Coronavirus Resource Center, a comprehensive collection of raw data and expert analysis from across a variety of disciplines.
With that project as a model, administrators hoped to replicate such collaborations in data science and artificial intelligence at scale across the entire university, Wirtz said. And the runaway success of the dashboard gave them a strong example to help sell the idea with donors.
Representatives for the university declined to say how much funding is going toward the new Data Science and AI Institute, nor have they identified its sources. Wirtz declined to name an anonymous donor who put forth unspecified funds for the institute, but hinted it is “unmatched in scale.”
Details about the institute itself suggest the university’s overall investment is sizable. The project includes the construction of two new facilities on the Homewood campus and the hiring of 110 new faculty and professors.
Officials have also launched an international search for a permanent director to head the institute. In the meantime, KT Ramesh, the university’s senior advisor for AI, and Rama Chellappa, a pioneer in the field of AI, are serving as interim co-directors.
The institute, they said, is designed to centralize a myriad of existing data science and artificial intelligence projects across the university. Existing programs and initiatives including the Johns Hopkins AI-X Foundry, the Institute for Data Intensive Engineering and Science, the Bloomberg Center for Government Excellence, the Malone Center for Engineering in Healthcare and the Center for Language and Speech Processing will contribute to the new institute.
“We had a lot of suburbs,” Chellappa said of such programs before the hub’s creation. “We didn’t have a downtown.”
A centralized hub means greater efficiency, standardized practices and stronger safeguards — all of which is bolstered by the diversity of expertise among institute-affiliated faculty.
“We as an institution are focused on doing things right, rather than doing things fast,” Ramesh said. “Taking the time to have the right people in the room … to engage in identifying both opportunity and risk, addressing how we can grow, how we can manage risk — all of this is really important to us.”
While 80 of the new faculty positions will fall under the Whiting School of Engineering, the remaining 30 will be Bloomberg Distinguished Professors. Those scholars are hired for their expertise in multiple disciplines and straddle two or more schools, divisions or departments. Officials hope a diverse range of experts will foster collaboration and creativity among the institutes’ projects.
New academic programs affiliated with the institute could become available as early as next fall, Ramesh said. The university’s investment is already catching the attention of academics around the country.
“It’s really exciting the way Hopkins is focusing on the future and trying to think about how to be a part of the big questions. And AI is a big question,” said incoming professor Louis Hyman, who will begin teaching at the university in January.
In Hyman’s current job at Cornell University, the labor, economics and business historian has experimented with artificial intelligence in his research. Hyman and a colleague painstakingly analyzed more than 300,000 letters written by Samuel Gompers, the first president of the American Federation of Labor. Earlier in the spring, he tinkered with ChatGPT to write computer code that could process images of the letters and translate them into text.
“That’s an incredibly trivial problem for a computer scientist, but for people like me, someone who is a humanist, it’s pretty amazing,” Hyman said. “Only you know what your problems and chokepoints are in your workflow.”
Although Hyman’s position at the university isn’t directly affiliated with the institute, he is already planning a course next semester on artificial intelligence for history. And he’s excited about opportunities the institute could bring for researchers across all disciplines.
Officials say the initiative will be a success if it leads to major discoveries, tools and inventions. And they hope the work of the institute may one day help shape public policies for the benefit of the country.
To Wirtz, the creation of a hub is most exciting because it won’t pigeonhole the university into a single subfield of artificial intelligence. There are a seemingly endless number of applications for the technology — from writing to gaming to medicine and public health.
The university may have attracted world renown for its COVID dashboard, but Hopkins is now building a broader reputation for multifaceted advancements in data science and artificial intelligence.
“We’re not a one-trick pony,” Wirtz said.