The early to mid-2020s were notable for the advancement of artificial intelligence technology, commonly shortened to A.I. The innovation of the modern language-model comes with a new wave of tools readily available to the public, such as a feature to communicate with an A.I chat-bot and ask any questions you wish, an image generation tool to create images at the user’s will, and even a video generation tool allowing you to create videos of any scene you want at the push of a button. The future is now, as they say!
Let’s start with common uses of A.I. that have wormed their way into near-daily use. Most Google searches result in an A.I. overview that explains the topic or answer you’re looking for. Social media apps like Instagram and X will give summaries and explanations for content you find on the app. Image generation tools like ChatGPT can create images of people to the user’s will, such as if they were in the 1980s, sporting fashionable clothing and styled hair suited for the decade. Content creation apps like YouTube allow users to re-imagine videos at the press of a button.
For example, a video I was watching on the platform asked me if I wanted to turn the background of the video into “candyland.” In another instance, the A.I. tool asked if I wanted there to be raccoons running across the screen.
Why is this option necessary, let alone added to the platform in the first place? Do we need an A.I. tool to simplify every action for us? Do we need every action to be made convenient and low-effort? When A.I. first became popular just a few years ago, it was marketed as being able to help researchers with discovering breakthroughs, finding the cures for previously incurable diseases, and pushing the limits on what the human race can ever think to comprehend about our world.
Instead, the main uses I see of A.I. today seem to be generating posters with copy-pasted styles, uninspired music, and artificial artwork, taking jobs away from actual creators who have spent years developing their craft, only for a “convenient” tool to copy the work of real people.
The future isn’t as bright as the convenience and efficiency of A.I. tools makes it out to be. For every flyer that a grassroots organization generates for an upcoming event, someone else is generating a video that purposely reinforces harmful stereotypes of a minority community. For every mom-and-pop shop generating a menu for their products, someone else is generating explicit content using the faces and identities of unwilling people. Is the convenience worth the price of harming innocent people?
To truly understand some of the consequences these A.I. tools have on our society, take into account the amount of water used and processed in A.I. data centers when cooling their servers and generating electricity. Sam Altman, CEO of OpenAI, says that the average ChatGPT query uses approximately 0.32 millimeters of water per use. However, some research, such as the research reported on by Forbes, estimates water usage in the range of less than two millimeters to over 150 millimeters for a single query. The difference in water usage typically depends on the language model being used, the length of the A.I. chat-bot’s response, the hardware, cooling systems, power source, and the location of the data center.
Miguel Yañez-Barnuevo, in an article from the Environmental and Energy Study Institute (EESI), states the following. “Large data centers can consume up to 5 million gallons per day, equivalent to the water use of a town populated by 10,000 to 50,000 people.” Carla D. Walker and Ian Goldsmith from the World Resources Institute talk about the water use of facilities further. “Mid-sized facilities can use up to 300,000 gallons of water a day, while large facilities can consume as much as 5 million gallons daily — comparable to what a small town uses.” Additionally, on top of the water demands, these data centers require huge amounts of energy, as described by Walker and Goldsmith. “A single modern AI data center can use as much power as 100,000 homes; many of the larger ones now being built are expected to consume up to 20 times that amount.”
The harm doesn’t stop there, as the many image and video generation tools made possible by generative A.I. allow people to create explicit and harmful content. In a recent case, as reported on by Ground News, a federal appeals court ruled that the First Amendment of the U.S. constitution protects possession of AI-generated child sexual abuse images, otherwise known as Child Sexual Abuse Material (CSAM). According to Ground News, Judge John Lee ruled that the possession of A.I.-generated CSAM is protected under the First Amendment as the material contains generated “depictions” rather than real children.
However, A.I. doesn’t just “generate” material. When images or videos are generated, real content from all over the internet are pulled into the amalgamation of the final product. Real photos and videos of CSAM, in this case, are used in generating these images and videos. When the court rules that A.I.-generated CSAM is only a depiction of children, rather than real children, it ignores the real children that the abusive material came from. There’s a troubling precedent that this ruling is setting, combined with the increasing number of data centers, and people losing their jobs to A.I. How many people will be harmed before our lawmakers draw the line?
As tech companies continue to push the use of A.I. into daily life, it’s ultimately clear that A.I. is becoming less and less favorable to the public. Data centers and generative A.I. have been affecting more and more people across the country through putting people out of jobs, risking clean drinking water, and allowing explicit material to be made of others without their consent. No matter how many people voice their disapproval of the concept, it’s clear that the companies and CEOs that control and fund these programs don’t have our best interests at heart, and are trying to inflate the worth of A.I. to their shareholders, essentially creating a “bubble” that resembles an echo chamber of approval of continued A.I. advancement from the voices that are creating these widely unwanted A.I. tools in the first place.
Being a Marine and Environmental Biology and Policy major, I’m worried about the consequences of unchecked generative A.I. on the people and on the environment. I have friends who specialize in art and creative design that are losing job opportunities due to generative A.I. being the “convenient” option over authentic, original work. Our individuality as people is what makes us human. Anything we do on our own, even if flawed, gives us room to grow and improve. If we’re always letting A.I. do the work for us, creating for us, talking for us, we lose what makes us human over time, and let an artificial personality fill the gaps.
I believe the A.I. bubble needs to pop.
