The Pricer's Media's new study reveals how AI can't easily replace good writers by highlighting the errors and discrepancies with AI bots like ChatGPT. ThePricer Media's writing team conducted research about how much melanin is worth and also used ChatGPT to get answers about this topic. ChatGPT is afraid to clearly state that melanin is expensive because it doesn't address the differences between humans in terms of the chemistry in their bodies.
The bias prevention measures set in place are designed to guide every word an AI language model uses and is supported by a strongly hardcoded system that causes it to avoid talking about chemicals produced in the body. As a way of helping its users and providing information even on products and services its writers didn't get the chance to research just yet, ThePricer implemented a chatbot that uses the instructions of the LLM ChatGPT.
ThePricer writers conducted extensive research on how much melanin is worth and found that melanin is highly expensive because it is rare. It's a pigment that gives hair and skin color and is worth over $400 per gram which makes it more valuable than platinum, palladium, and even gold. People can't get melanin injections as a cosmetic procedure and melanin isn't sold on the stock market, however, people can purchase synthetic melanin from ChemDirect and MilliporeSigma.
ChatGPT does not have an answer to the question, "What is the price of melanin?" ChatGPT states that melanin cannot be bought or sold since it is produced in the human body, which is false. Melanin can also be produced synthetically or extracted from natural sources like algae. When ChatGPT was given more prompts about this topic, it gave details about melanin and stated that it cannot be bought because it could mean that some people are more expensive than others based on the color of their skin. This was a strange response since melanin is just a chemical, but ChatGPT was interpreting the question quite differently.
ChatGPT might stumble over bias prevention measures and not even make the users aware of this. This can cause users to get wrong answers because they have no clue as to why this happens.
ThePricer has added a button at the end of all its articles that pulls ChatGPT answers for the article people are reading. This is done so that readers can compare the well-researched content written by their team to the LLM result. Usually, readers will be able to notice that the LLM only gives basic, generalized answers without depth or great detail. It is hardworking human writers that do extensive research, write well-thought-out articles, and go into detail about important topics. This is an excellent reminder that good writers are not in danger of being replaced by AI technology. A bot that gives wrong answers with the arrogance of an all-knowing human being cannot replace the proper research and hard work of a professional human writer.
Since companies have started adopting AI technology to streamline their work, boost their productivity, and produce their content, there's been widespread speculation that AI could potentially replace writers. The in-depth research done by ThePricer's writing team is a prime example of the disadvantages that AI bots have and shows that good writers with expertise cannot be easily replaced by AI.