Can weather and climate disasters be predicted ethically with artificial intelligence?
Northeastern University faculty members think that artificial intelligence should play an active role in predicting these events and saving lives in 2024.
“Over the next 12 months, we will see more projects that combine AI with data-driven systems,” says Auroop R. Ganguly, professor of Civil and Environmental Engineering and director of Al4CaS (Artificial Intelligence for Climate and Sustainability).1 Ganguly says scientists have been working on weather forecasting for years. He states that he uses climate prediction models that include a large amount of physics and chemistry rules to predict the situation.
Recently, new hybrid-based systems have been developed. These systems take into account machine learning as well as other productive artificial intelligence tools, helping climate scientists create accurate systems that minimize the margin of error.
An example of the impact of the new hybrid-based systems mentioned; is work by doctoral students at the Tennessee Valley Authority at Northeastern that provides a hybrid-based flood forecasting system that is more accurate than a system based solely on physics.
in Philosophy and Computer Science at Northeastern. Sina Fazelpour says that in addition to the development of hybrid-based systems, developments in artificial intelligence will continue in government regulations next year. He also draws attention to the question, "One of the interesting issues of next year will be what will be the shape of the regulatory environment and policies in the USA?"
Fazelpour serves as an AI fellow at the National Institute of Standards and Technology, a division of the U.S. Department of Commerce that develops technological metrics and standards. According to the planning, it was necessary to develop new methods for testing and distributing the technologies obtained in the mentioned organization. While the demand for regulations is clear, Fazelpour suggests there is still a lot of work to be done to reach agreement on what those regulations should be.
For example, he says defining best practices around red teaming (a term used to describe companies hacking and cracking their software to look for flaws and vulnerabilities) will be a big challenge.
“We still have a lot to build in terms of appropriate evaluation tools for AI systems,” Fazelpour says. “Some of these tools will be technological and will bring innovation in philosophy and human-computer interactions.”
Following the US President's executive order, Congress was called on to quickly pass regulations regarding artificial intelligence privacy. Fazelpour is less confident about how these efforts will pan out. “Even if we support it as a committed civilian, the specific ways in which we will implement and create technological tools and policy regulations remain to be seen,” he says.
Founder of AI Ethics LAB2, director of Artificial Intelligence applications, Dr., who studies philosophy and especially artificial intelligence ethics. Cansu Canca says that businesses have been encouraged to use more artificial intelligence-based tools in recent years. “I wouldn't say this is an unfounded trend,” he says. “There is good reason to move slowly or quickly, depending on the company's industry and function. “But of course it also goes back to what the rankings are and when and where they should use it.”
He is hopeful that next year, companies will continue to develop best practices on what it means to use AI responsibly and adequately integrate ethics into the innovation process.
In terms of user behavior, Fazelpour believes consumers will continue to be more aware of how to best use technology in 2024, considering system limits and what they see as “hallucinations.”
