Marking the start of a new era: An overview of generative AI

in Features by

When individuals think of AI, what often comes to mind are manipulated images and text generators. This version of AI is known as generative artificial intelligence (gen AI) and is used when creating new content like audio, code, simulations, videos, images and that 3-page explanation of what a chromosome is. Gen AI has been around since the 1960s, but has seen significant advancements in the past decade. From the rise of ChatGPT to the widespread online implementation of company AI chatbots, gen AI has become one of the most widely-accessible branches of artificial intelligence.

According to the 2024 McKinsey Global Survey of AI, gen AI is becoming increasingly implemented and accepted by individuals and corporations. Last year marked the beginning of gen AI’s potential as a resource for businesses, with 65% of survey respondents globally stating their organizations regularly used this branch of artificial intelligence. Overall adoption surged to an all-time high of 72% in 2024, as found by a separate McKinsey survey. With this increased interest, changes in schools and workplaces have become noticeable, causing ethical concerns related to misinformation and privacy to arise.

Generative AI tools are trained on vast image and text databases from varying sources including the internet. ChatGPT, for example, sources from the OpenAI company platform, meaning its data comes from publically-available datasets; anywhere from web pages to published journals is fair game. Therefore, not only is that question about chromosomes being answered with an article from the National Library of Medicine, but it’s also being answered with a Reddit response under r/explainlikeimfive. Although this is an extreme oversimplification of the model, the basics remain the same. Gen AI commonly sources from any publically-available database, regardless of its quality or factuality. 

Additionally, gen AI poses privacy issues for individuals such as artists. For example, in 2022, Tennessee-based artist Kelly McKernan, known for their distinctive blend of Art Nouveau, science fiction and an acid-colored palette, found their name featured in a list of prompts to feed AI in order to generate “Lord of the Rings”-esque art. They were not alone: several other artists’ names were also listed. 

“It just got weird at that point. It [the art] was starting to look pretty accurate, a little infringe-y,” McKernan said in an interview for “The New Yorker.” Gen AI art generators draw from artists indirectly as well, often sourcing images posted online as a base for what it creates. Commonly, these artists do not give permission for their artwork to be used, and they do not receive any monetary compensation for the use of their art.

In contrast, one of the most well-known art-and-AI debates arose during the 2022 Colorado State Art Fair.  The first-place winner of this competition was Jason Allen. His piece “Théâtre D’opéra Spatial” depicted a woman looking into a circular portal with two other individuals in robes near her. The point of the controversy was that the artwork was made entirely using gen AI in a platform called Midjourney.

The competition’s judges were unaware that AI had been used but said they would have awarded him first place regardless. This provoked a mixed response across the internet. Although gen AI had been used in the creation of art pieces previously, “Théâtre D’opéra Spatial” created a controversy and discussion about ethics and the creation of art. This drew attention to the question: what is art? Should a generative AI artwork be classified in the same category as pieces created by the human hand?

Since 2022, the use of gen AI has continued to grow in the fields of business, art and the humanities. However, the ethical controversy continues.

Can you tell human from bot? Scan the code above and try out these online games which challenge you to decide if you are talking to a human or gen AI. The original Humanornot.ai was created by AI21 labs and can be seen on the left. The independently-developed Humanornot.so game is on the right. Try them out and compare your skill. (Graphic/Olivia Chung)