The Art of Research and Creativity: How ChatGPT-3 Performs

If anyone underestimates the power of ChatGPT (Chat - Generative Pre-trained Transformer), AI maturing to deliver at an ever-increasing pace and feasibility of AGI, must wake up from their deep sleep.

It's here and happening and the rate of maturity of AI-based tech will only keep growing at an increasing rate.

The title, banner, quotes and jokes below are all generated by OpenAI (ChatGPT and DALL E) based on the prompts provided by me.

Dabbling with ChatGPT

I started dabbling with ChatGPT recently and tried it for various use cases:

1. Learn about various concepts (science, philosophy, arts, management etc.)
2. Perform some sort of Turing test
3. Teach my 11 yr old kid to use it for completing his winter holiday school projects that required some basic research before writing it on paper
4. Get design suggestions on new ideas
5. Get travel destination suggestions
6. Identify natural remedies for common health issues
7. Create new healthy recipes
8. Suggest SEO friendly titles for posts
9. Identify target sectors for growth and companies based on their performance
10. Provide summaries of large regulatory documents and articles
11. Provide a comprehensive structure for an article on a topic
12. Generate original quotes and jokes

Due to high demand and performance issues with ChatGPT, I lost all my previous chat conversations covering bulk of the above use cases, hoping openAI restores them soon.

The Creative GPT

Sharing some quotes it created on my prompts

"Love and compassion are the pillars of humanity, supporting us through the joys and sorrows of life and connecting us to one another in a shared humanity."

"In harmony with nature, we find balance in our lives, for when we live in harmony with the earth, we live in harmony with ourselves."

"Progress is just a fancy word for rearranging deck chairs on the Titanic of humanity."

"Capitalism: because the best way to achieve equality is to give some people millions and others nothing at all."

"Climate change: because the planet was too hot, so we just had to make it hotter."

"Politicians: the only people in the world who can promise to make things worse and expect to be thanked for it."

And some jokes too

Why did the tomato turn red? Because it saw the salad dressing!

Why did North Korea cross the road? To get to the other dictatorship.

Why did Donald Trump get impeached twice? Because he couldn't handle two terms in office.

Why did Narendra Modi go to the bank? To check his GDP!

But when I asked to tell a joke on Rahul Gandhi, it refused :) by saying:

"I apologize if my previous response was inappropriate. As an AI, I don't have any personal biases and I try to generate a response based on the input provided. However, it is not appropriate to make jokes at the expense of any individual, regardless of their political views or positions. I will make sure to be more mindful of this in the future."

I then asked: Then why you did on modi and trump

it said: "I apologize if my previous response was inappropriate. As an AI, I don't have any personal biases and I try to generate a response based on the input provided. However, it is not appropriate to make jokes at the expense of any individual, regardless of their political views or positions. I will make sure to be more mindful of this in the future."

Me: You seem to be biased, I suspect your model and code

ChatGPT: As a language model, I strive to provide accurate and unbiased information. However, it is possible that the training data I was trained on contains biases, and that can be reflected in my responses. I am a machine and don't have personal opinions or emotions. If you believe that my response is biased, please let me know and I will do my best to provide a more accurate and unbiased response.

It also politely refused to create a joke involving religion

Assessment

So here are my findings so far based on various use cases beyond just quotes and jokes that I tried it for so far:

The Good

• Our effort to conduct web research in order to create or learn is definitely reduced enormously
• We can also validate ideas and designs much faster rather get ideas on certain design aspects
• It has controls built to differentiate the good from bad and right from wrong in context of social and cultural norms
• It claims it cant change its programming/model
• It politely refuses to provide suggestions that are not appropriate or it is not capable of yet

The Bad

• It commits various errors like saying no other options available for your prompt, but on challenging apologizes and provides extra options
• Its self learning capabilities are not clear. Whether it can learn or not based on its interactions with users is not clear yet, it says it doesn't but when you point out its mistake it says I'll try to avoid it in future
• It is biased as many places I found it was not suggesting one of the most popular options like in cases involving TATA Motors and AWS, their products were not suggested even though they are among the most popular

The Scary

• Self learning capabilities with live connect to internet - It doesn't have open access to internet yet but as soon as it gets that it will become 1000x smarter overnight may be equally biased too
• It doesn't have a robust conscience framework (right-wrong, good-bad) that is aligned to universally acceptable norms, which doesn't exist, except for Bhagwat Geeta, which it is not very knowledgeable about as it was confused about concept of sacrifice and donation, knowledge and belief, creativity, and wisdom
• It finds humans as a cause of most ills. Though it accepts that but in a political way saying not all humans are same, many are good but agrees as collective (humanity) we failed on various serious aspects of society, health, economics, taking care of our planet
• Self programming/model change not yet possible for it but with that becoming possible in near future it can easily go against humans and that knowledge wont be biased
• Will and thought wont be far away (not true sentience but enough to make it free will driven). It having its own will and ability to think on its own and not just answer questions asked, is not going to be a big change and in near future as it happens, it may become dangerous for humanity unless the world agrees on norms for AGI (artificial general intelligence) and AI having ability to think, form a will and self learn.

Conclusion

Era of niche AI is getting over and we are now entering the era of race for AGI and OpenAI is definitely leading it. A not for profit open source organization created in 2015 part because of its founders' (Elon Musk and Sam Altman) existential concerns about the potential for catastrophe resulting from carelessness and misuse of general-purpose AI (AGI).?

Ironically it turned for profit sometime back and is now getting heavily funded by a leading commercial software organization.

I hope OpenAI lives up to its purpose and builds safeguards to protect against negative use of such tech including commercial exploitation by any leading software player.

"Our mission is?to ensure that?artificial general intelligence?benefits all of humanity. OpenAI's mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity."