AI tools have generated a media frenzy since ChatGPT hit the scene. The general message people are receiving is that these tools generate creative material, but the use of the word creative is misleading. These bots, once out of an infancy of being spoon-fed curated datasets and human reinforcement, prowl the internet or other designated hunting grounds vacuuming up words, images, music, and more that humans created. Basically, these bots are mining for patterns and relationships. For example, how often and how closely A is related to B.
Once a bot has “learned” enough, it is finally able to simulate creativity. To do this, they operate at supercomputer speed, recombining atoms of human thought and artistry using probability to guess what the next word in a sentence should be or the next brushstroke in a painting or the next note in a musical score. The result is believable mimicry of human creativity. This is an oversimplification and generalization of the code mechanics involved. The point is that it is nothing but mechanics.
These machines possess none of the hallmarks of true creativity and originality: inspiration, creative leaps, intuition, emotions, subconscious, intelligence, sentience, or spontaneity. They lack even the most rudimentary understanding of what they are doing and outputting. Let alone any understanding of the human experience, which is the wellspring of all art.
Many news reports show the quality of what is produced by this AI mimicry is sometimes eerily remarkable and sometimes not exactly mind blowing. When compared to what a skilled human can do, let alone a Picasso, a John Lennon, or a Hemmingway, these bots fall far short. So perhaps it is premature to raise the specter of concern that these AI tools are pillaging the creative heart of humanity just yet, but that day will come if we encourage it.
For now, perhaps the real question is why would someone want to read literature “written” by an AI that is mimicking human creativity by outputting a mashup of millions of authors. Reading a novel or essay about human life “written” by a lump of lifeless silicon is the same thing as dating a robot. It misses the entire point of dating or reading, which is forging a connection between two humans.
Why would someone want to listen to music created by an AI that mimics musicians instead of music that has real soul? Why would someone want to gaze into the depths of a painting created by an AI that mimics emotions instead of gazing at a masterpiece created by someone who has loved and lost the greatest love of their life and expressed those emotions in their art?
Consumption of this AI mimicry is like swallowing a placebo for reality. At its essence, the purpose of the arts is to communicate the human experience, what it means to be alive. Creating a bot that mimics artists is not the sincerest form of flattery, but sincerely missing the point entirely.
It is one thing to create AI prototypes that mimic human creativity or emotions in a lab for the purposes of scientific research. It is quite another thing to release these bots into the wild, often completely untested for safety or ramification. Far too many people already equate the mad ramblings on social media to journalistic quality news. What will happen to the truth when these bots start making believable word salad out of everything for everyone?
AI tools that consume and analyze mountains of data to find patterns—causes and effects—will someday soon become invaluable gifts to humanity. These kinds of AI applications hold the promise of everything from early cancer detection to assisting in the quest for cures. AI robotics systems that automate complex processes, such as driving a car by deep “learning” from example, will also be vastly beneficial to humanity. These are very different AI applications than those that are engineered to mimic human creativity or even sentience.
This blind race to achieve AI pseudo-sentience and mimicry of human creativity would be harmless except for a few things.
One: The fallout from every AI advancement in human mimicry dehumanizes us all a little more.
Two: The use of creativity mimicking AI tools is the same as condoning forgery. All the humans whose art was vacuumed up and predigested to produce this mimicry are victims of a sneak thief in the night. Furthermore, isn’t building a high-speed machine that mimics the brush strokes or writing style of an artist attempted theft of their very soul and, in a way, our souls too?
Three: The use of these creativity mimicking AI tools will dumb us all down until, ultimately, on some glorious distant day, we will all have become nothing more than receptacles of computer generated gibberish. Garbage in, Garbage out.
This dumbing down will likely begin at school and carry on all through life. Students using an AI chatbot to write essays is not the same as moving from slide rules to calculators, as some suggest. Writing an essay requires thinking deeply about that topic and working out one’s ideas. It is the very foundation of understanding and the exercise of human intelligence. Comparing chatbots to calculators is nonsensical. Using chatbots is the elimination of using one’s brain intellectually and creatively. What we do not use, we lose.
The AI genie is already out of the bottle and coming for us all, but not in the way most sci-fi books and movies portray. There is a philosophical concept that one cannot transcend oneself. Since we are incapable of fully understanding our thinking processes, we are incapable of creating a Frankensteinian computer that can truly think as we do. So there will be no singularity that gives birth to malevolent sentient machines. Instead, we will have something completely different, alien, and no more alive than a lump of coal or a loaded gun.
We need societal guardrails. AI tools should be limited to things like robotics or data analysis. AI should not be used to create forgeries of human creativity. If we do not limit some of the uses of AI, then the dreaded AI apocalypse will come, but it will not be some sci-fi Terminator Skynet war of humanity versus machines. It will be the slow drip drip drip loss of our humanity, and once our humanity is diminished enough, so too will our very purpose for living.
References:
https://hbr.org/2018/01/artificial-intelligence-for-the-real-world
https://hbr.org/2022/03/customer-experience-in-the-age-of-ai
https://iopscience.iop.org/article/10.1088/1742-6596/1575/1/012192/pdf
https://www.inc.com/hillel-fuld/5-ways-to-use-chatgpt-in-your-workflow.html
https://www.kungfu.ai/blog-post/we-used-chatgpt-to-figure-out-how-businesses-can-use-chatgpt
https://www.usatoday.com/story/tech/2023/01/27/chatgpt-buzzfeed-ai/11129947002/
https://www.searchenginejournal.com/is-chatgpt-use-of-web-content-fair/477558/
https://www.washingtonpost.com/technology/2023/01/27/chatgpt-google-meta/
https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html
https://www.digitaltrends.com/computing/the-6-biggest-problems-with-chatgpt-right-now/
https://www.washingtonpost.com/technology/2023/01/27/chatgpt-google-meta/
Integration:
https://cadabra.studio/blog/about-gpt-chat/
Marketing issues:
https://www.bloomberg.com/opinion/articles/2023-01-26/chatgpt-is-not-much-of-a-pitch-robot#xj4y7vzkg
Writers:
https://hechingerreport.org/i-wish-a-bot-were-smart-enough-to-write-this-column/
Replacing workers:
Resistance:
https://www.sciencedirect.com/science/article/abs/pii/S0306457322003582
https://iopscience.iop.org/article/10.1088/1742-6596/1575/1/012192/pdf
ChatBot Customer Satisfaction:
https://www.sciencedirect.com/science/article/pii/S0969698922001527
https://www.information-age.com/evolution-artificial-intelligence-customer-exer-10025/
Types of AI & R&D:
https://arstechnica.com/gadgets/2023/01/the-generative-ai-revolution-has-begun-how-did-we-get-here/
How do Chatbots work:
https://www.context.news/ai/what-is-chatgpt-and-will-it-steal-our-jobs
https://www.sciencefocus.com/future-technology/gpt-3/
https://towardsdatascience.com/how-chatgpt-works-the-models-behind-the-bot-1ce5fca96286
https://towardsdatascience.com/generative-chatbots-using-the-seq2seq-model-d411c8738ab5
https://pub.towardsai.net/chatgpt-how-does-it-work-internally-e0b3e23601a1
Not creative: