What’s Next With Artificial Intelligence?

Today, AI technology can pull up the weather on command and answer your online shopping questions. But how does a machine understand what we’re saying? The answer lies in the field of natural language processing, which uses advanced AI algorithms to analyze large amounts of text and develop word sequence associations. When someone talks or types a question, the machine draws from these associations to examine the context and predict the best word sequence to answer the question.

“Researchers are impressed by what AI can do with language,” Dr. Ruihong Huang said. “But the underlying model is not how humans think. AI doesn’t really know what you’re talking about because it cannot reason.” As the director of Texas A&M’s Natural Language Processing Group, Huang studies how AI can better mimic human reasoning when analyzing language for tasks like identifying social media posts about local disasters or detecting media bias in news articles.

While the human-like AI programs from the movies are still just a fantasy, Huang hopes that advances in natural language processing will produce machines that can be more accurate and help people and companies with simple yet time-consuming tasks. “There have been amazing developments with AI,” she added, “but we still have a long way to go.”
 

As with any technological development, AI is not without its problems. Though Siri offering the wrong information might be a simple inconvenience, mistakes become more serious as AI technology undertakes greater responsibilities, like identifying promising job candidates or qualified federal aid applicants. Such systems could also be trained on incomplete data, causing unintentional bias toward different groups.

For Hannah Bloch-Wehba, addressing these problems is essential as more governments adopt AI decision-making technology. “Often, it’s not very transparent how an AI arrived at a decision, which leads to less public accountability,” she explained.

Currently, AI development and use has little regulation, and lawmakers are still in the early stages of addressing potential disadvantages. “In the past few decades, technology regulation has been largely permissive to encourage innovation,” Bloch-Wehba said. “We’ve never confronted technology quite like AI before, so the question of the right legal approach is still wide open.” As society continues to incorporate AI in new ways, lawmakers will have to grapple with how to protect values like privacy and equality to make the technology better for everyone.
 

What is free will? What does it mean to have agency or moral responsibility? What is creativity? These are questions philosophers have discussed for centuries. Now, they have another question: Can AI technology ever possess these characteristics?

Not for a while at least, said Dr. Martin Peterson. “There is a fundamental difference between how a human brain and an AI system works,” he said. “If we consider more than just the output and focus on how it was produced, we wouldn’t say that current AI technology can truly think or be creative.”

When considering current and future AI developments, Peterson recommends focusing less on aspects the average American can’t control and more on using it in a way that aligns with your values. “For example, even if an AI could write all my philosophy papers, I’m not sure I’d want it to because I believe it’s important for me to write those papers to develop my philosophical skills,” he explained. “Even though AI systems can write papers or compose music, it’s still impressive when humans do these things. AI creates material by drawing on existing papers or music, but humans can create something entirely new.”