�㽶��ý

The AI of the future is here.

An automated robot vacuum, video calls with friends and another with your boss, smart homes with endless capabilities and wearable technology.

What sounds like today’s norm is the thing of tv cartoons past.

"The Jetsons" cartoon, airing originally in the 1960s, showed a family of four, along with their robot dog, and how they used artificial intelligence (AI) tools to make their everyday tasks and lives simpler.

Jump to 2024, and we are engaging with AI tools as part of our daily lives to accomplish the same goals. We are remotely controlling the lights in our homes, tracking our heart rates and sleep patterns from a watch, asking voice assistants to answer simple questions and using navigation tools to help us with traffic.

In the last year, the popularity of generative AI, which utilizes data sets to create originally generated content in the forms of text, images, and audio has risen significantly. The most popular and often talked about AI program is ChatGPT — an AI chatbot tool utilizing foundational large language models, such as GPT-4, that generates dialogue.

“Everyone is talking about ChatGPT and for good reason — it has redefined the standards for artificial intelligence and shown how machines can learn the complexities of human language and communicate with you through a chatbot. The experience of communicating with the chatbot is no different from us chatting with a friend via text messaging,” said Shweta Kailani, M.S., MBA, director for the Center of Teaching and Learning, who was an early adaptor of the technology in 2022. “I'm passionate about technology and new innovations pertaining to higher education — so I had to look at this new platform and ask, ‘Is this a fad? Is it something that's going to stay, to become a part of our daily lives and be ubiquitous and accessible to all?’”

The future is now

The capabilities behind generative AI go further and are much broader than most users are aware of.

“AI’s ability to perform cognitive functions associated with the human mind are ever expanding —  from the perceiving of content, the ability to learn and adjust, to problem solving and creativity. Everyone assumes it is the future when we speak about AI and what it can do — but the future is now and is here,” said Bjarne Berg, Ph.D., mathematics and computer science professor, who is an internationally recognized expert in business intelligence. “We are seeing advances in technology that people expect to be years away when in actuality it is occurring now and being put to use. For instance, when you go on Netflix and are looking for a particular movie, but do not know the actors or the title, the meta tags are not giving you the result you want. But Google is working on an image search within the movies themselves. So, you say, ‘Show me all the movies where a woman in a black skirt slides on a banana peel’ and then all movies where that occurs appear. The search is occurring within the movies themselves and not just how they are tagged.”

It feels unbelievable that every movie could be referenced using one detail or a particular scene. However, according to Berg, it is real and happening.

“That's what people said in 1996 when we first had a search engine that searched all the Internet. Think about today, you search the whole Internet at sub-second speed, everything that's ever been written by billions of people, and we take it for granted,” said Berg.

What other Jetsons-style imaging technologies are also on the horizon? According to Berg it is a blend between video gaming and movies.

“Today, video gaming generates more revenue than movies in the U.S. And most of the things that you see in the video game, such as the non-active background character, people popping up, or buildings and frames — instead of game designers having to draw everything out, they're using AI to create randomization within the game,” said Berg.

But what does that look like?

If you play the game five times, it may be slightly different every time. A building might have changed in height or color; the person walking across the street might be a woman in a red dress and then next time is a man in a black suit.

“In gaming, the game is no longer designed, it is probabilistic. We're basically saying that every time you play the game it'll be slightly different, and therefore it keeps you interested. You don't even notice. A lot of times people play the game five times, they didn't notice that a lot of the backgrounds and the non-actors are generated and are constantly changing,” said Berg.

Movie studios such as Pixar and DreamWorks are creating films that also generate content that make it so that you might watch your favorite movie 15 times and each time, it's slightly different.

“In a computer-generated movie, you can do a random assignment. This character could either walk across the street; he can attack you; he can jump a tall building; he can bite you. And there's all these behaviors you can assign and randomly pick one every time. So even the people who made the movie have no idea how it's going to end. So instead of having a movie that today is curated, cut, edited, and fixed to always be the same movie, it becomes new and fresh every time,” said Berg. “In other words, you may be willing to see the same movie 20, 30, 40, 50 times because you don't really know how it's going to end, and it changes every time. Isn't that kind of cool?”

As a form of visual storytelling, these evolving forms of AI are generating headlines that could write themselves.

“The attractiveness of these AI programs that can do everything — be the writer, editor, a curator of facts — it is exciting,” said Kailani, who has held multiple workshops on campus for faculty looking to learn more about generative AI and how to integrate it into their classrooms. “AI in general has a big role to play, especially when it comes to things relating to time efficiency. For example, if I'm working on an essay or research paper as a student and need help from the Writing Center to get guidance on writing style, grammar, sentence construction or organization, I need to wait till the next day when I can make an appointment with them to seek help. But here you have generative AI, and you can just ask it to proofread for you, and boom, you get a pretty decent response, which again, you have to still go back and check behind it — but it does the bulk of the heavy lifting.”

What’s the problem?

As incredible technology breakthroughs continue with AI — they also bring about challenges and problems. The knowledge base for ChatGPT’s version 3.5, the free version most accessible to Internet users, is limited to information shared in 2021 and before, which can often present inaccurate or false information that sounds not only plausible but factual.

“Say you ask the platform to tell you about the medical standards for a colonoscopy. It will share what it believes is the most up-to-date and populated answer,” said Kailani. “But the standards were changed within the last year and were not accessible to you since it was behind the paywall for the paid version 4.”

As digital citizens consuming information through these platforms, the discernment between fact and the generative AI hallucinations is needed since information is neatly packaged for users, making it hard to detect the misinformation or assumptions that the tool has created.

“If I go into the program and ask it to write me a story about Dr. Bjarne Berg at �㽶��ý, it will start by pulling all the information available on the Internet and compiling from there. Having done this exercise in class, it crafted the following: Dr. Berg started his career in computer science — and right there it got it wrong. I do not have a degree in computer science, but given the frequency of my name being associated with computer science, it automatically assumed I had a degree in the field when really I started my career in the Navy,” said Berg. 

These generative tools' main goal is to condense large amounts of content and data into responses that are helpful. The accuracy of the information generated is a secondary goal.

“AI generated tools are a great place to start when it comes to compiling data, text and information — but you cannot take it all for fact — just like the hallucination about my degree, it just scraped together information it found,” said Berg. “So the burden then falls on users, and in our case students, to do the extra legwork to ensure the content they asked to be generated is factual. But also — computer generated content does not care about being polite, racist, sexist or using inappropriate language. Since AI is learning as we code and teach it, it is going to discriminate.”

As more faculty are implanting the use of AI into their courses, students are expected to cite AI as a source — just as you would any other website, journal or book.

“It is up to us as educators to educate our students about how to approach AI tools in a way that is ethical — ethical in how we teach it, how we use it for research, how our students approach it for their usage and in ensuring academic integrity is present,” Kailani shared. “We are encouraging students to use critical thinking evaluation skills to figure out the weak spots — what did it miss? Was there a bias in the response? AI generated text often has biases and doesn’t look at a topic or prompt from multiple perspectives. Often times, it provides a very superficial overview of the topic and doesn’t cover all perspectives or an in-depth critical analysis of the topic/text.”

As a senior computer science and cybersecurity major, Ringo Nguyen ’24 has familiarized himself with many AI platforms including ChatGPT, Google’s Gemini and Microsoft’s Copilot.

“Generative AI is a trend that is not going away anytime soon, so it is important to have a working knowledge of multiple platforms,” shared Nguyen. “As we familiarize ourselves with new and emerging tools, we need to see them as one source that needs fact-checking and to be supplemental — not the only answer.”

His advice to students and others who are looking at AI generated content and want to use it?

“There are always pros and cons to using new technologies — especially those that are on trend. But developing a deeper understanding of the technologies will provide you with so much knowledge and benefit you in the end. Embracing innovation with an open mindset will enable you to adapt and thrive in today’s rapidly changing world.” said Nguyen. “These technologies and tools were created for human beings by human beings. So, at the end of the day, we are creating a future benefitting everyone.”