By SHANNON O. WELLS
The Dec. 1 meeting of the Senate Computing and Information Technology Committee included discussions of such complex tools as artificial intelligence-based software to detect plagiarism and the rollout of Claude for Education at Pitt (see related story), but started off with the more pedestrian notion of a communication gap regarding recent, campus-wide cybersecurity training.
Committee Chair Angela Zack said comments she received from “various people” who didn’t realize the cyber training-related email sent to the Pitt community earlier this fall was legitimate, have led to additional clarity from Pitt Digital, which led the training module. Specifically, an email message received from Senate Council indicated the cyber training invitation contained “kind of mixed signals.”
“The notification email itself ‘contained multiple fonts, colors and links to unknown URLs,’” she reported from the Senate message. “‘It looked exactly like a phishing attempt, effectively teaching users to ignore the red flags the training is trying to spot.’”
The message also expressed the perception that training instructions effectively lowered security defenses.
“To access the training, users are instructed to disable their pop-up blockers, which contradicts standard advice to keep browser security settings high,” Zack reported, adding that the third point of the email regarded content quality. “‘The quiz questions contain confusing or technical poor advice, such as implying that a secure password only needs to be several characters long.’”
After the Senate email was forwarded to Pitt Digital, Zack said improvements in training — based on a recent reminder she received — are evident.
The new email explicitly tells users to type pi.tt/securitytraining into their browser or search for "security training" on my.pitt.edu or to click the link Dashboard | KnowBe4. “It also referenced the University policy AO 48 (University Information Security), and explained why pop-up blockers need to be adjusted, because the training is hosted by the partner, which is NoviFour,” Zack said.
“It's a small detail, these changes, but it makes the email look less like a phishing attempt,” she said. “It was good to see that that feedback was working.”
Image duplication detection
Zack went on to discuss potential advocacy for Proofig AI software, an AI-based tool that checks research manuscripts for image duplication before submission to publications. She referenced a Nov. 5 Faculty Assembly meeting discussion of new policies and procedures for research misconduct.
“It was discussed that our faculty should have access to tools to catch honest mistakes before they become reputational issues,” she said.
Since the Senate Council asked the computing committee to advocate for purchasing an enterprise license for Proofig AI software, Zack said she confirmed that Bill Yates, vice chancellor for research protections, “supports the use of this software,” and has discussed supporting it with the Health Sciences Library System, where Zack works.
“I think where the bottleneck is now is there's currently no funding for an enterprise license, and there are departments that are buying individual licenses,” she said. “But that's insufficient and creates equity issues with the use of the software.”
Describing the software’s specific function, Zack said it’s her understanding that Proofig AI checks an image to make sure it’s not something copied from somewhere else or is an “incorrect, AI-generated image.”
“There may be more research that we have to do into that,” she said, noting she asked Yates to attend the next committee meeting in January “to give us more information. And if he has time, he will do that.”
Kenny Doty, associate director of technology in the Swanson School of Engineering, said while he’s not opposed to incorporating the software, he’s concerned that requiring it could inhibit timely submission of manuscripts.
“Erroneous references are rampant out there, and we want to publish the best papers possible — for ourselves and the institute,” he said. “We all have tons and tons of images. … A lot of times papers are being submitted by people in my group at 2 a.m. on Saturday. So if it's a tool that they could use, but (must) be given approval for something — and then some administrator has to then say, ‘OK, you can submit your paper’ — I think those are two different things.”
Mark Henderson, vice chancellor and chief information officer, suggested continuing the discussion. Coming back with “some kind of formal recommendation to the broader Senate, might be in order,” he said. “Then it can go forward.”
“We're happy to go out and negotiate enterprise licenses for the institution — that's what we do — and get the very best price. The challenge is, unless there is a designated funding source for this, that will be problematic. …
“Another way would be if the deans found favor with this and are willing to fund it,” Henderson added.
If it’s determined to be an enterprise application, he added that “Pitt Digital would likely support it. … We would just build that into whatever the overall cost and cost recovery mechanism would be.”
Klaus Libertus, teaching associate professor in the Department of Psychology, said he’d like more information about how the software can be used in other disciplines.
“This is (based on) chemistry and biology, and I can see how that works there, but is it broadly applicable to everybody who publishes? That would be interesting to hear what they say about that, because their website has slices and other plots that you see in biology that does not necessarily apply to everyone.”
Storage challenge
In other topics, Henderson offered an update on the status of OneDrive data storage in the wake of Microsoft announcing that in 2026, any entity that goes beyond a particular threshold will have to pay for it.
“And we are way over this stipulated threshold,” he said, noting the issue has been shared with the Dean's Council and Academic Leadership Council. “At our current clip, it would increase our cost by over a million dollars a year, so we're working on a number of fronts. … We've analyzed this every way that one can, and we have very, very good data that we are sharing with the deans to have them advise us on (how to) address this.
“We're not telling anybody what to do,” he said. However, “if we don't have retention guidelines or retention policies for information, we can beat this down, but sure as we're talking, it'll creep back up over time, and we'll be back in the same boat. We think, through our analysis, there are some recommendations that can be taken.”
These include opportunities to move things out of OneDrive into other storage, and still have it archived, “where you can bring information back or go after it as an option to — if it's deemed appropriate.” Responsibility centers also could put caps on what people can keep “so that we don't breach whatever our limit is and incur costs” while also considering specific needs and circumstances.
Noting that he thinks the OneDrive policy goes into effect later in 2026, Henderson said “we hope to be way ahead of it in terms of our way forward before then.”
The next computing committee meeting is scheduled for 1 p.m. Jan. 26.
Shannon O. Wells is a writer for the University Times. Reach him at shannonw@pitt.edu.
Have a story idea or news to share? Share it with the University Times.