It is the biggest event since the Apollo moon landing in 1969 — no, this even rivals the Industrial Revolution. Some people could not stop praising the revolutionary technology revealed on March 15. And that technology is ChatGPT version 4, the latest upgrade of the interactive artificial intelligence chatbot developed by the artificial intelligence company OpenAI.
ChatGPT has gradually improved since the release of its first version in November 2022. Not only does ChatGPT-4 provide outstanding results, but its impressive suite of creative capabilities may send chills down one's spine.
One of its best features is that anyone with an internet connection can use it for free. Normally, AI software necessitates the purchase of additional equipment as well as documentation and payment for use. But this is not the case with ChatGPT.
Spine-Chillingly Smart
Here is a closer look at what makes ChatGPT-4 so astounding.
For example, ChatGPT-4 passed the mock bar exam in the United States and even scored in the top 10% of examinees. (Version 3.5 scored in the bottom 10% and failed.) It can also perform at human levels in other professional and academic settings.
When instructed, ChatGPT-4 can also write novels, essays, scripts, and poems. Not only can it write text, but it can also generate responses based on provided images. For example, when shown images of flour, eggs, and milk and asked what one can do with them, it will inform you that you have unlimited options, like making pancakes or waffles.
It can even explain the humor of a funny image, write lyrics, and compose music.
But unsurprisingly, it also faces a plethora of problems.
For example, ChatGPT-3.5 would frequently (and confidently) provide irrelevant answers to questions. A person who entered their name was informed that they were a failed politician, despite not sharing a name with a politician.
Another inquired about nearby restaurant recommendations, but only one of the five suggestions provided by the chatbot actually existed.
Will AI Replace Knowledge Workers?
However, problems like these will probably be resolved in the blink of an eye. At times, the chatbot's responses are aggressive or even discriminatory, but these issues, too, will likely be addressed in the near future.
On a more serious note, people are concerned that AI will replace knowledge workers. I'm too afraid to try, but let's say I tell the chatbot to "write a convincing argument piece about ChatGPT-4 in the style of Kumiko Takeuchi." I'm confident it could put together a seemingly credible argument.
Academic performance will undoubtedly suffer as well. The majority of students admit to using artificial intelligence to help them with homework and reports. In New York, students and teachers are banned from using ChatGPT. Naturally, it will cause dysfunction in academic disciplines.
Torrents of fake news will be generated, and public opinion will take a strange turn as human opinion is mixed up with the opinions expressed by AI. Public opinion is already being manipulated to some extent, but this will threaten democracy even more severely.
In response, OpenAI's creators are improving their tools for determining whether something is a human or an AI creation. The goal is to provide a countermeasure to information manipulation, misguided public opinion, and fraud in education and academia.
When discussing breakthrough AI innovations, the topic of technological singularity always comes up. Will the singularity happen, and when? Inventor and philosopher Ray Kurzweil presented this concept in his 2005 book The Singularity Is Near: When Humans Transcend Biology. It denotes the hypothetical point at which AI outperforms human intelligence or has a profound impact on human life.
Parenting Wisdom
Kurzweil predicted that the singularity would occur in 2045. However, given recent advances in AI over the last several years, we can speculate that this day could arrive sooner.
In response, Sam Altman, one of the OpenAI founders, explains that the singularity is certainly coming. He says that the question is whether AI would develop a will of its own.
Elon Musk is one of the people who founded OpenAI alongside Altman in 2015. However, he left the company three years after it was founded. He was concerned about the viral phenomenon of information spreading indefinitely once it appeared on the internet. On December 4, 2022, he tweeted, "ChatGPT is scary good. We are not far from dangerously strong AI."
At any rate, now that we have gotten this far, it is not the time to debate whether AI will ever develop a will of its own. We should assume that it will almost certainly happen.
In such a situation, we have no choice but to respond with the knowledge we have accumulated as humans. We should use our parenting expertise, in particular, to grow, handle, and pacify AI. Because after all, humans are the birth parents of AI.
RELATED:
- Robotics Bringing New Comfort for Lonely Seniors
- AI Senbei: We Tried an Onigiri Rice Cracker Developed with Artificial Intelligence
- Level-4 Self-Driving Cars to Hit Japanese Streets for the First Time
(Read the article in Japanese.)
Author: Kumiko Takeuchi
Kumiko Takeuchi is an animal behavioral scientist and essayist.
