Maryland researchers working to correct potential biases in artificial intelligence
BALTIMORE - Artificial intelligence can make your life easier.
But could an AI model cause harm?
There's critical research being done in Maryland is to consider equity and make sure AI systems are not biased.
Artificial intelligence has a huge impact as it's now intertwined with almost every facet of our lives, relying mostly on worldly influences.
Along with other researchers, University of Maryland, Baltimore County Professor of Information Systems James Foulds, and senior Shaniah Reece are working to correct potential biases.
"It's going to pick up on the issues in our society and reflect them and maybe even amplify them," Foulds said.
"So, algorithmic fairness, what we seek to achieve, is to correct those bias within the data," Reece said.
Researchers say AI still has a long way to go and should only be used in collaboration with human intelligence to help account for checks and balances.
"AI algorithms take data from our society so that gets baked into what they learn which could lead to unfair predictions that lead to unfair decisions that affect you," Foulds said.
AI complies data, statistics and other research, then forms its own conclusion.
It aids in the job selection process, it can determine if a person is eligible for bail or parole and so much more, but sometimes those outcomes can be detrimental.
"It is very important that we consider how technology and AI can impact those decisions and especially because we know that minority and underrepresented groups are disproportionately impacted by it," Reece said.
Diverse researchers are joining efforts to catch potential biases and advocating for transparency across the board.
"Transparency would be a big move with regards to helping to resolve some of these issues because it would leave the process open to scrutiny which can help with identifying biases before they occur," Reece said.
The White House has introduced a blueprint for an "AI Bill of Rights," identifying principles that should guide the design, use, and deployment of automated systems to protect the American public.
Related coverage
- 'AI can be a friend or a foe:' As we become more reliant on artificial intelligence, focus should be on balance, expert says
- Maryland researchers are working to curb biases, inequity in Artificial Intelligence
- Artificial intelligence being used to answer questions about sea level rise
- Meet ChatGPT, the artificial intelligence chat bot that could take your job
- Cutting-edge UMBC research uses artificial intelligence and robots to assist national security