The Human Factor in AI with Tamilla Triantoro
Tamilla Triantoro, associate professor of business analytics and information systems at Quinnipiac University, discusses the human factor in artificial intelligence.
- By Upside Staff
- January 25, 2024
In this recent “Speaking of Data” podcast, Tamilla Triantoro discusses the human factors of artificial intelligence, including the latest developments and their ethical implications. Triantoro is associate professor of business analytics and information systems at Quinnipiac University and speaker at the upcoming TDWI Modern Data Leaders Summit. [Editor’s note: Speaker quotations have been edited for length and clarity.]
As an academic, Triantoro’s research focuses on human/AI collaboration and specifically on the ways AI can augment human capabilities in various capacities. This involves training and tuning AI systems for specific tasks as well as examining the impact on the decision-making process and work overall.
“One of the more exciting developments,” she began, “is with multimodal AI -- that is, AI that can hear, see, and speak. If AI can now interact with voice, text, and visual data, there’s the potential for it to be able to interact with the world around us -- the people, buildings, public transportation, and so on -- a network of AI agents interacting with other agents.”
Triantoro brought up a recent article in the journal Nature about Google DeepMind’s work in the area of scientific discovery.
“They used graph neural networks to generate potential new materials, which is usually a tremendously time-consuming task when scientists have to manually combine various elements in the lab to see what works and what doesn’t.” She explained that the system has already generated more than 2 million potential materials -- any one of which could be used to build new homes or create more efficient superconductors -- of which scientists have tested 700.
She continued, “I’m not too scared of generative AI, but I am curious to see what will happen to humans. Will we still be as creative as we were or will we rely mainly on the output of our systems?” Triantoro explained that from a worker’s point of view, their job is generally broken down into tasks they like and those they don’t, while from the employer’s perspective, that same job is broken down into tasks that generate profit and those that are considered cost centers.
“If companies can figure out the place where profit-generating tasks overlap with the tasks the employee likes doing, they can create a more dynamic, motivated, and efficient workforce,” she said. “In our research, we’re trying to figure out where AI can contribute to that.”
She noted that although AI can contribute to any task, for those tasks that are pure, highly researched creative tasks, people outperformed AI on a regular basis.
All this, however, raised the question of ethical implications of AI in the workplace.
“Work is being done to establish ethical frameworks,” Triantoro said, “even if only in response to recent legislation passed in the EU and in the works elsewhere around the world. However, there are some things the technology can do that people just aren’t ready for yet.” She raised the subject of automated hiring and firing -- something the technology is perfectly capable of doing. However, from a human point of view, there’s a potential for serious public backlash and reputational risk should such a thing become known to the public. In addition, she noted, any system that treats employees as mere data points runs the risk of winding up with an alienated and mistrustful workforce.
“AI tools are being introduced into the workplace,” she said, “but what’s important is how it’s being done. It’s great when AI performs tasks we don’t like, but when AI starts taking over the tasks that employees take pride in or think of as core to their professional identity, problems can arise. Companies should cultivate partnership between people and AI to clearly show the benefits to humans.”
She also recommended that employees should begin exploring AI tools such as ChatGPT or Bart if they haven’t already. She suggested that regular interaction with AI tools would improve their comfort level for when similar tools are ultimately part of their jobs.
[Editor’s note: Dr. Triantoro is presenting a session entitled "Expert Best Practices: Navigating the Human Factor in the AI-Powered Workplace" at TDWI's Modern Data Leader's Summit in Las Vegas (February 19-21, 2024). For information about the Summit's full agenda, visit TDWI.org.]