In the Age of AI, Humans Will Still Be in Charge

GEOINT educators discuss the educational needs of the future

Education-Training-Panel

Geospatial intelligence in an AI-enabled world is about adaptation and trust. But it’s not about accepting that the human element is being moved to the rear of the bus by a machine.

“It’s about combining machines with human processes,” said Dr. Todd S. Bacastow, a panelist in Sunday’s “Education vs. Training: GEOINT in an AI-enabled World” session as part of GEOINT Foreword.

Bacastow, a professor of practice for geospatial intelligence at Penn State University, added: “Unless we’re going to have machines direct machines, humans are going to have to be involved and we’re going to have to have humans direct machines to augment what we do.”
 

Bacastow and fellow panelists offered their perspectives on technological advances that have some geospatial practitioners concerned.

Increasingly, the discipline is likely to involve “life-long learning,” said Dr. Camelia Kantor, Vice President of Academic Affairs for USGIF. It will also include the necessary ability to adapt to change.

“How do we react to innovation and change every time we have a new revolutionary technology?” Kantor continued. “How can we react when we know there’s a change coming?”

Before universities establish curricula to teach evolving geospatial technology, including AI and machine learning, academia needs to accept that change is occurring quickly and will continue to do so.

“It is a little hard to keep up,” said U.S. Army Col. (Ret.) Steven D. Fleming, Ph.D., professor of practice in the Spatial Sciences Institute and the Institute for Creative Technologies at the University of Southern California. “Sometimes, one step behind is the best you can do. In many cases, we are doing all that we can do to stay up with the educational and training requirement.”

One method educators can use flies in the face of traditional higher education, which requires lengthy approval timelines for new courses and curricula. Flexibility is paramount to incorporate evolving technology.

“I encourage academics that are in the business of doing any kind of education [to be flexible],” Fleming said. “They should say, ‘I’m going to accept the fact that this discipline is dynamic, and I’m going to address change through the unknown, looking over the horizon, and you folks are going to have to trust me.’”

Trust them to do what, exactly? To answer the question, “What are the competencies that people who do geospatial intelligence should have in the AI environment?” Bacastow elaborated.

“One thing we can predict is that we’re going to need to train the AI systems,” Bacastow said. “We need to teach people to be experts at training them.”

People also are necessary to interpret AI results, to apply critical thinking to the human-machine team, and to apply prescriptive analytics to amplify anomalies. They will also be needed to cope with ongoing technological developments.

Those people will be best taught in modern settings. The era of PowerPoint in support of a lecture does not fit the dynamic nature of the geospatial intelligence profession.

“We’re going to have to move into a much more realistic training environment, or education environment,” Bacastow said. “We’re going to have to put the student in very difficult situations. That means a large variety of real-world training scenarios. That’s where the human will interact with the machine and gain confidence in the ability to use it.”

Teaching will change, but so will learning, according to Fleming.

“Alvin Toffler said, ‘The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn,’” Fleming quoted. “Something that was right when you learned it, then wrong today, can become right tomorrow. You have to be a dynamic learner.” 

USGIF Volunteer Spotlight: Todd S. Bacastow, Ph.D.

Todd S. Bacastow, Ph.D., is a longtime supporter of USGIF and the GEOINT profession at large, and currently leads the Foundation’s Academic Planning Committee

Bacastow was teaching at Penn State shortly after 9/11 and around the same time the term geospatial intelligence (GEOINT) was coined, when his then colleague, David DiBiase, attended the first GEOINT Symposium.

“He came back and said, you’ve got to see this,” Bacastow recalled. “It was a natural fit in away. In a larger context, [getting involved with USGIF] was my way of continuing to serve after I retired from the Army.”

As a result, Penn State became one of the first universities to earn USGIF Collegiate Accreditation and to grant academic GEOINT Certificates.

Current priorities for the Academic Planning Committee are looking ahead to understand the academic needs of the GEOINT discipline, how USGIF can support those needs, and how to move the discipline into the next decade. This includes the planning of a new academic GEOINT summit USGIF plans to host in September 2020.

Bacastow’s overarching goal for the committee is to “take the intellectual lead on the discipline.” To do so will require diversity, he continued, urging academics who are just learning about GEOINT to get involved.

“This is a way they can learn the discipline, and we value their participation because they will broaden the perspective,” Bacastow said, adding he hopes to reach other markets such as business and public safety.

As the discipline matures it is starting to see a second generation of professionals with a stronger grasp on AI, machine learning, and deep learning, according to Bacastow. 

“That’s in a real sense where the future lies—it’s in the new people, it’s in that new generation coming up,” he concluded.

  • To learn more about USGIF’s Academic Planning Committee and educational initiatives, contact USGIF’s Dr. Camelia Kantor at camelia.kantor@usgif.org.

, ,

The Science of Certification

Developing performance standards for the GEOINT community

, , ,

Mike Campanelli: Forward Thinking

Young Professional Mike Campanelli is the new USGIF board member

,

Visualizing the Invisible

GMU pioneers a new approach to harvesting GEOINT