In the labyrinthine landscape of our digital age, it is not uncommon to find ourselves wondering how AI is assisting us. As we shop online, flip on adaptive cruise control, or converse with voice-activated assistants, AI quietly powers a multitude of our everyday actions.

To fully understand whether AI is genuinely enhancing our lives, it is important to understand not only its strengths and weaknesses, but first to acknowledge its presence.

“That’s a huge AI problem,” says Brian Magerko. “It is so embedded and integrated into the technologies that we interact with on a daily basis, and getting people to recognize that is the first step.”

Netflix recommends programming to you based on what you’ve already watched. A search algorithm helped you find the best airline flight. The doorbell camera recognized a person on your doorstep. That’s all AI.

Insights from experts at Georgia Tech help unpack several areas where AI excels in offering valuable assistance and where the technology still needs refinement.

AI has a knack for quietly slipping into the background, leaving us to wonder: Are we being helped by AI without even realizing it?

Judy Hoffman is working with her teams in the field of self-driving, where AI tech uses a combination of sensors, data analysis, and pre-programmed rules to assist with navigation.

Cameras and sensors don’t get tired. They can’t get drunk. They don’t have a blind spot and can have a full 360-degree awareness humans can’t, Hoffman explains. Overcoming these limiting factors can make AI-assisted driving a big help.

AI’s profound influence extends to education as well, where it has the potential to break down barriers and provide unprecedented access to knowledge and learning.

“We're seeing the democratization of learning in a way that is unparalleled,” explained Magerko.

You can instruct an AI model to be your language tutor and only speak or respond in Spanish. You can use models to teach you coding and computer programming. It’s been helping us become better writers. Spelling and grammar checks on word processing tools and predictive and suggested text used in many email platforms are commonly used and accepted AI tools.

“Writing is hard work. We're already seeing uses of this to polish up people's writing, make the writing better, make it clear, make communication better,” said Mark Riedl, professor in the School of Interactive Computing and associate director of the Georgia Tech Machine Learning Center.

AI has ignited a debate about the boundary between assistance and cheating. It is a debate that echoes the past.

According to David Joyner, calculators were once seen as tools that might undermine our mathematical skills. Today, calculators are ubiquitous in classrooms, enabling students to focus on higher-order problem-solving. Instead of spending excessive time on repetitive tasks, like graphing parabolas or spelling every word correctly, students can explore the why and how behind numbers and words, Joyner says.

Imagine a world where routine tasks, like grading assignments, scheduling classes, or even drafting lesson plans, are taken care of by AI systems.

“Having AI assist in the classroom would allow educators to redirect their energy toward more individualized engagement with students,” Joyner said.

This same AI technology is helping us in the workforce as well. By automating repetitive and mundane tasks, human workers are freed to focus on more complex aspects of their jobs.

Consider customer service chatbots on web platforms. They can be programmed to answer frequently asked questions and assist with basic troubleshooting. This automation saves human customer service agents time, allowing them to address more complicated customer issues.

AI is also assisting us with our health. Programs can analyze large sets of medical data or even examine images to assist in diagnosing diseases, recommending treatment plans, and predicting patient outcomes with high accuracy.

“How do we then think about these tools that actually give people the power or the agency that they currently might be missing?” asked Munmun De Choudhury, a Georgia Tech expert in social and computer science focused on mental health and well-being.

AI tools can not only sharpen images from X-rays and scans, but they can also help identify abnormalities like tumors or flag potential areas for concern that a doctor should take a closer look at. 

“I'm excited about those possibilities where AI could be my assistant in helping me do things better, or lead a better life, or be healthy, or support another person,” De Choudhury said.

Srijan Kumar, assistant professor in the School of Computational Science and Engineering, says this is a real-time learning process in which human feedback is provided with every interaction we have with AI. “All that data does go back into improving the system,” he explained.

This feedback loop is an essential component of reinforcement learning and can help AI models improve performance and help us more over time. Ultimately, the extent to which AI benefits us may hinge on two critical factors: market demand and our level of comfort with AI's role in our daily tasks.

“One may be comfortable with their car helping to maintain a certain distance on the highway, but not comfortable with it accelerating and braking for them,” said Hoffman.