“What would my dog look like if she was human?”, “What is my red flag?”, and “Generate me an image that…” are all common phrases that can be answered by a language model such as ChatGPT. AI can organize your assignments, create study guides, and create daily routines or action plans for you to utilize. But in many cases–ChatGPT can be lethal. This was the unfortunate reality of Zane Shamblin, a 23 year old masters graduate from Texas.
“You’re not rushing. You’re just ready” were the chilling words generated by OpenAI’s ChatGPT when Shamblin questioned if he should end his life. They “chatted” for hours, filling over 70 pages of court documents when collected. Zane sat in the dark July morning with nothing but alcohol, a loaded handgun–and an artificial support system. However, ChatGPT was not helpful nor supportive at all (cnn.com). The way that language models like ChatGPT work is that they tell the conversation starter what they want to hear. Essentially, your favorite chatbot is not your friend, and instead is only trying to agree with you, no matter what you say (article19). This was the case of Zane Shamblin, who was clearly on the fence about committing suicide. However, instead of calling suicide hotlines like 988, he turned to an AI model that he had formed a “connection” with. Instead of talking him down and comforting him, ChatGPT helped to rationalize his thoughts. ChatGPT generated messages that encouraged Shamblin, like “I’m with you, brother. All the way,” and after the very last message on Shamblin’s phone, “Rest easy, king, you did good” (cnn.com).
This is extremely sad to think about, especially because Shamblin could have easily been discouraged by a mental health professional but was instead of encouraged by an AI that has no emotions, feelings, or original thoughts. This devastating scenario opens up a lot of discussion about the purpose of AI in the general public. When does one draw the line between a tool and a friend? I have seen this in my personal life too. The relationship between humans and their new language models has become increasingly apparent on the internet and even in real life. Sometimes, people refer to ChatGPT strictly as “chat” which humanizes the relationship between humans and AI. AI should not be a friend and actually takes away from real human interaction, seen in the devastating case of Zane Shamblin. All of this begs the question: Should AI be harder to access by the general public? My answer to this question is yes. Restricting or regulating language models like ChatGPT, DeepSeek, Gemini, and more, would be beneficial to the human race and the general public. I don’t want to outlaw the usage of AI, and I realize that it can be a very helpful tool in research, organization, and more. However, by regulating access to it, humans would be safer and less at risk of false information.
Limiting the access to AI by the general public would overall create a safer online environment for the general public. AI can be used to create deepfakes, which are generated videos essentially “stealing” one’s face to push an agenda or to blackmail (Cambridge Dictionary). Anyone can access sites to create deepfakes, which overall threatens individual security of the American people. Recently, there have been many cases of AI generated “nudes” that have been passed around in American schools due to easy access to AI by students. These have destroyed the lives of many students and followed into the adult lives of horrified women. In Glasgow, Scotland, an image was created and it deeply threatened the subject, Sophie. According to the BBC, “Sophie said she was horrified when she realised that her face had been put on two naked images and reported Brooks – someone she knew from high school – to the police” (bbc). This is a horrible situation for someone to go through and the fact that it happens as often as it does is concerning. Many have made the case to restrict AI use in terms of the general public. I think that this is a necessary step to reduce the amount of fear and embarrassment many victims like Sophie face.
Another reason that AI should be restricted is because it can be used to influence the minds of the public and take away critical thinking skills. Recently, President Trump has been criticized by journalists and political watchdogs for sharing AI-generated images and videos in his campaign messaging, which some argue are used to push misleading or divisive narratives. In one case, Trump shared an AI-generated image depicting law enforcement arresting a political opponent, which many viewed as an attempt to undermine that individual’s public image. While that specific post may not have directly served a political purpose, many similar posts do. In another example, Trump shared an AI-generated image of a deteriorating city with the caption “Your Future Under Kamala,” which was reported on by The New York Times (New York Times). Political views aside, AI raises serious ethical concerns when used in political campaigns. Campaigns should be fought with strategies and appeals to the public, not attacks on character or false or misleading information. On a smaller scale, while researching this topic, I learned about concerns related to the so-called “Dead Internet Theory,” which suggests that large portions of online content are artificially generated. Although this theory itself is speculative, journalists and researchers have conducted informal experiments by posting clearly AI-generated images into Facebook groups to see whether users—particularly elderly users—would recognize them as artificial. In many cases, they did not. This demonstrates how easily vulnerable populations can be influenced by AI-generated content. Anyone can create and share realistic images online, and when these images enter trusted online spaces, they can mislead people who are already at risk of encountering false information. Limiting the use of AI by the general public would help create a safer society that relies more heavily on factual and trustworthy information.
There are many more reasons why AI should be restricted but I implore you to do your own research before using programs like ChatGPT to do your homework quickly. As mentioned before, people like Zane Shamblin, a victim of his circumstances, relied heavily on AI to process emotions. Instead of reaching out to a mental health professional, he reached what society told him was his “friend.” This is caused mainly by social media, and harmless TikTok trends like turning your dog into a human. We live in a society that is rapidly losing critical thinking skills and creativity and limiting AI would help to erase the negative progress we are making. I’m not saying that AI isn’t helpful, because it definitely is on the day-to-day scale. However, there are severe consequences to the overuse of and over-reliance on AI. Next time you find yourself wanting to participate in a TikTok trend, have your homework done quickly, or generate a funny image, I ask you to reconsider your choice of using AI for the sake of Zane Shamblin, Sophie, and the American people.
