Baltimore Illustrator John de Campos isn’t happy that AI is being employed as a tool to produce images for marketing uses replacing artists whose actual experiences are reflected into their art. (Andrea Hsu/NPR)

Baltimore Illustrator John de Campos was irate when he found out that some of his work from the beginning was used to teach an AI chatbotwithout his consent.

“It’s disgusting,” he says.

In the last few years AI-powered programs such as Midjourney as well as DALL-E have allowed anyone to create sophisticated images in just two clicks on the keyboard.

For de Campos, it’s an outrage.

“The possibility that expression and art by humans is at risk and could be on the verge of disappearing is incredibly scary for me.” He says.

However, de Campos, who is hoping to make money as a game designer is finding ChatGPT to be an useful tool in the marketing of his games through social media.

“I’ll declare that these are the attributes of the game we’re selling. Consider all of this information and break it into fifteen words or less. I’d like to see five different versions that you can use to sell this item on Instagram,” he says.

De Campos is aware of the skepticism that this creates and says he’s trying decide if he wants to separate himself from any AI software.

“I find myself a bit enthralled by the artistic side, but not as much about the text aspect however I’m slowly finding out why,” he says.

John de Campos says he worked for 50 to 60 hours drawing illustrations for his newest board game Black Mold, which he describes as a horror-themed survival escape. A few of his rivals in the game industry are using AI to create art. (Andrea Hsu/NPR)

One year after the introduction of ChatGPT people from all sorts of professions are trying to decide what to do with AI. Views on the technology differ widely, and there is no consensus on what jobs should and cannot be given to robots.

The most efficient writing assistant

Within Michigan, Ethan Kissel has turned to AI to assist him in his work of producing TV commercials for local businesses.

“It’s great for spitballing concepts,” he says.

The past was when Kissel used to spend an hour or more creating an hour-long or 30 second script, the most difficult portion being the taglinethe most memorable phrase. Kissel discovered that ChatGPT can create dozens of taglines in two seconds.

“Most all of these are likely garbage,” he says. “But you take a little from one, and a couple of words from another and put them all in one and then you’ll end up with something that you actually sort would like.”

Kissel is able to easily envision a time when copywriters will be eliminated and voice actors who provide narration. The company he works for already has an AI tool that can correct pronunciations in the event that they’re not able to finish their work on time.

At present, however, Kissel is less worried about his work. Along with the writing of scripts and stories, he edits and shoots video, and has meetings with clients. Being a jack-of-all trades is, according to him, some security.

“I do not think it’s the same as it is in ‘the present moment. However, it’s one we have to consider and prepare in advance,” the politician says.

Change is fast approaching.

In all different fields, there are clues of what’s to come. Newspapers have employed AI to produce reviews of high school sporting games. Video game companies have been using AI to design new characters. Software developers use AI to create codes.

Karin Kimbrough, chief economist for LinkedIn which is part of Microsoft does not see AI becoming the norm for all jobs immediately. However, she believes AI is certain to change the way people are spending their time on their job.

“You may find yourself spending less time working on mundane activities,” she says. “You may devote more time to things… you require the human power of your skills as well as your empathy skills and ethical judgement, as well as your abilities to manage and directing people.”

In the end, Kimbrough says, the hopes are that AI can make humans more productive and efficient.

There are many mistakes to avoid on the path. The web is full of reports of AI chatbots that are able to deliver complete creations.

An New York lawyer was sanctioned this year after he was caught with fake cases in a lawsuit filed against an airline. The lawyer in court acknowledged that he’d utilized ChatGPT to research legal issues and didn’t bother to check the bot’s actions.

These kinds of cases help to see how reckless usage of AI could harm individuals and society.

Understanding the weaknesses and risks

Jeffrey Garcia, a program manager for a tech firm and has taken himself on himself to determine what AI excels at as well as what isn’t not just for fun, and to stay up to date with technological advances that could impact his future job.

The experiments he has conducted over the last year has provided him with an insight into the limitations of technology and its potential.

As a kid, Garcia was always frustrated by his inability of capturing the beautiful images he that he imagined in his head.

“I have a great love with art” He says. “But I don’t like it.”

In the spring of this year the entrepreneur wondered if he could launch an Etsy store, with a few items he had created using Midjourney? The first project he attempted was an old-fashioned poster of an iconic bird that he loves, the Baltimore oriole. Baltimore oriole.

The program presented a sophisticated picture of an oriole facing the Baltimore skyline, however there were some glitches. Garcia’s wife, who is a biologist, noticed an extra pair of Toes that the bird’s feet had. The skyline has some landmarks that are well-known, however it’s not a great fit to anyone who has a good understanding of the city.

Garcia hasn’t decided to sell one of his posters, but found it to be easy to sell an item, particularly given the speedy advancements AI tools for creating images have made. In the fall, he experimented using the recently released DALL-E 3 to create the identical poster, and he was able to create a better image, but it’s not entirely free of anatomical mistakes.

Jeffrey Garcia with an image of an Baltimore oriole he designed with DALL-E 3 in the month of October 2023. (Andrea Hsu/NPR)

“It’s improving however, it’s still not enough to meet the standards of a person who is aware of the situation,” he says.

These experiments have informed the way Garcia utilizes AI to work. Garcia sees ChatGPT as a non-sensical assistant that is able to create drafts of first drafts, however his work needs to be vetted and rewritten.

There are a few aspects of his work that aren’t yet his to give up. For instance, correspondence is among them.

“I am not comfortable leaving this thing I consider vital and deeply human to a computerized system,” he says.

Copyright 2023 Copyright NPR. To learn more, go to https://www.npr.org.