Is AI bad or just Scary?

Eleanor Roosevelt, whom some believe to be the first female President of the United States, once observed famously:

“Do what you feel in your heart to be right - for you'll be criticized anyway. You'll be damned if you do, and damned if you don't.”

I am not particularly enamored with her choice of words, but the message is powerful. Her phraseology came back to me recently when I saw a news story in the New York Post regarding a major manufacturer and its attempt to engage Artificial Intelligence (AI) in support of the holiday spirit (and perhaps product sales).

This time of year, there is always some surge in advertisements celebrating emotions, spirit, and goodwill. It has traditionally been a season of celebration, and there have always been businesses that seek to attach good feelings to their services or products.

Some claim that the whole "Saint Nick" legend was largely shaped by a Clement Moore poem in 1822, which evolved to be "The Night Before Christmas." Find me someone who has never heard that. I dare you. Many would credit that as the beginning of "a Father Christmas (as) an enduring part of the Christmas tradition."

But, it was a manufacturer (the same one in the New York Post story) that "first ... executed a marketing campaign for Christmas with Santa." That was in 1931, almost a century ago, adapting visual images then recently created by "cartoonist Thomas Nast" and "artist Norman Rockwell." And it was the beginning of a veritable landslide of marketing and endearment.

The point has been driven home "many times, many ways" (The Christmas Song, Nat King Cole, Capitol 1946). In a nutshell: holiday spirit and Santa sell.

The New York Post reports that in 2024 this manufacturer "collaborated with three artificial intelligence studios." The result was various "advertisements that ... feature furry animals and" the company's products. There is apparently a theme of nostalgia as these pay homage to previous ad campaigns of yesteryear.

The story describes that these sequels "have received intense public backlash, with critics calling them "soulless." The manufacturer is accused of "cutting corners" and "sacrificing quality for time and money." The heart of those allegations is that the company used AI to generate these ad spots, "instead of hiring actors or animators."

The news story then devolves into a discussion of the cost of AI. There is even description of how much carbon dioxide might be generated in creating an AI ad. The Post notes that "the commercials quietly exhausted valuable resources and contributed to Earth's overheating by relying on AI." No, the author of the Post article is not John Greene Chandler, but some might think so. 

Ho, Ho, oh No. The article does not provide any comparison to humans in terms of the exhaust gas. But, it notes that "Generating 1,000 images using AI creates as much carbon dioxide as driving a gas-powered car 4.1 miles."

The U.S. Environmental Protection Agency (EPA) says "The average passenger vehicle emits about 400 grams of CO2 per mile." So, perhaps it is fair to say that the 4.1 miles cited by the news would be something less than 2,000 grams? That is 2 kilograms.

According to Slate.com, "The average human exhales about 2.3 pounds of carbon dioxide on an average day." For the sake of comparison, we have to reconcile the American anti-metric posture. According to the EPA, one pound of carbon dioxide is equal to .545 kilograms (a kilogram is one thousand grams). 2.3 pounds times .545 kilograms means the average human expels about 1.25 kilograms in a day.

So, for 1,000 images, the AI produces 2,000 grams of Co2. If a human-produced 500 images a day, the 1,000 human-made images would cause about 2,500 grams of C02. How many images a day is realistic for one human to produce? Would the 500 in a day be in 8 hours, so the human can go home? Or, would the comparison use the 24 hours that the AI will work without a break even for the restroom?

In an 8-hour day (without breaks), the 500-image prognostication would require the human to produce 63 images each hour, or just over one per minute. Is that realistic? I know I get up to stretch my legs once in a while throughout the day. I have also seen some artists work and that seem quite an adventurous pace. Is it fair to conclude that AI picture production is actually less damaging to the environment than human C02 exhalation? (If you subscribe to human climate change)  

The story further reflects the negative of AI. It suggests that the shortcomings of the output ("unnatural movements or anatomical inaccuracies") support the argument for more human involvement. It accuses the ad of exacerbating the planet's temperature. Essentially, there is the suggestion that this ad campaign just plain ruins Christmas.

But, in the end, AI is a novelty. No one has a monopoly on using AI. I know people who are using AI to great effect. They are composing music, producing videos, and even writing articles. They love it, and apparently, AI loves them too (I cannot get many of the AI tools to work, but I keep trying). 

On one hand, perhaps in the absence of AI, humans would be employed at significant C02 cost to produce the same or similar content my acquaintances are making. On the other hand, perhaps without AI that content just would not be produced? If people could not have AI compose them a theme song, would they hire a musician or band instead? Doubtful. More likely, they would do business with out a theme song.  

There is room for concern. There will be impacts of AI. There are inherent dangers in AI. There is potential for misuse and even abuse of AI. But the same can be said of many tools we have seen born, raised, and welcomed into our society (remember when the wheel came out, wow!?! What an uproar that caused). There will be bumps, bruises, and changes in days to come. Just remember "change is the only constant" in the world, and that alone is the one thing that will never change.

Prior posts on AI and Robotics

Will the Postal Service be our Model for Reform? (August 2014)
Attorneys Obsolete (December 2014)
How Will Attorneys (or any of us Adapt? (April 2015)
Salim Ismail and a Life-Changing Seminar (May 2015)
The Running Man from Pensacola, Florida (July 2015)
Will Revolution be Violent (October 2015)
Ross, AI, and the new Paradigm Coming (March 2016)
Chatbot Wins (June 2016)
Robotics and Innovation Back in the News (September 2016)
Universal Income - A Reality Coming? (November 2016)
Artificial Intelligence in Our World (January 2017)
Another AI Invasion, Meritocracy? (January 2017)
Strong Back Days are History (February 2017)
Nero May be Fiddling (April 2017)
The Coming Automation (November 2017)
Tech is Changing Work (November 2018)
Hallucinating Technology (January 2019)
Inadvertently Creating Delay and Making Work (May 2019)
Artificial Intelligence Surveillance (August 2020)
Robot in the News (October 2021)
Safety is Coming (March 2022)

Metadata and Makeup (May 2022)
Long Term Solutions (June 2022)
Intelligence (November 2022)
You're Only Human (May 2023).
AI and the Latest (June 2023)
Mamma Always Said (June 2023)
AI and the Coming Regulation (September 2023)
AI Incognito (December 2023)
The Grinch (January 2024)
AI in Your Hand (April 2024)
AI and DAN (July 2024)
AI is a Tool (October 2024)

Rights for the Toaster (October 2024)
Everybody Wake Up! (October 2024)
First What is it? (November 2024)
X-Files or Poltergeist? (November 2024)

Is Gartner Helpful on AI? (December 2024)

The Eeeeyew AI Says What? (December 2024)

Is AI bad or just Scary? (December 2024)