Artificial Intelligence - Where is it all going?

Photo by Tara Winstead (Pexels)

I know that I already wrote a similar article about this topic two years ago (two years? Time flies anyways) as one of my first articles published on this blog. But as mentioned in my end-of-year recap article I did mention that I had the intention of rewriting some old articles with a new perspective and I do believe many things have changed since I last wrote that article. 

For instance, AI is more so on the rise than ever. At least more people are talking about it on the news and much to my surprise in day-to-day life too. It is just interesting to see what AI is now and what it potentially could be in the future.

And please do not worry, this article was not written by an AI of any sort. This was written by me because I wanted to write about my perspective and not let an AI write about my perspective. Even though it would have been funny to let an AI write about itself.

First of all, I would like to define the topic of today’s article as always to get everyone on the same page. And funnily enough, I did spend some time trying to define artificial intelligence and all I came up with is artificial intelligence is “an intelligence made by humans to be smarter than humans.”

However, that definition is overly simplified in my opinion that is why I also found the dictionary definition which says that AI is “the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.”

Now looking at the dictionary definition of artificial intelligence, I do believe that the definition would be changed in the near future since I believe the potential of AI can go far beyond what human intelligence can such as visual perception, speech recognition, decision-making and translation between languages.

Somewhere in the innovation part of my mind, I believe that AI can become so much more than just another human with human intelligence. I believe that AI is capable of so much more than our current intelligence can imagine. Somehow, I imagine that AI eventually will discover another ability which goes far beyond human comprehension. Since I am not that up to date on current technology news, it could have already happened. 

Back in 2017, I believe, Facebook made two AI named Alice and Bob whose main function was supposed to negotiate but apparently, they managed to create their own language while still fulfilling their main functions as negotiators. Using their own language, they came to a conclusion fast. In other words, the language changed for a specific reason, so the conclusions have come quickly and efficiently. 

While the reasons Facebook abandoned that project are still debated among the media as I have read in many sources and they either say that the AI research team were afraid that Alice and Bob would turn into something uncontrollable together or the AI were not fulfilling their purpose correctly which is negotiating. 

Whatever the reason maybe the Facebook AI were shut down and AI in general disappeared in the public eye for a while the researchers are still working on AI advancements. I mean there was AI news here and there but mostly not so much.

According to Google Trends searches the searches mentioning AI skyrocketed at the start of 2022. So, between 2017 with Facebook AI and 2022 with the hype of open AI, there was not much of AI to go around. However, now AI is all there is to see.

Everything on platforms ranging from creative platforms to learning platforms has some sort of AI integrated into it. The AI on those platforms are not as advanced as the ones in development but they help out in small ways and do their jobs perfectly.

Then you have the other AI more in the public’s eyes. The advanced ones would write an essay for your or create a picture just from simple descriptions. The ones who would step in when you do not have time to do your own research or just generally do not want to do the task.

I am talking about the AI we can interact with to make our lives easier. The biggest player currently in the game is OpenAI with their ChatGPT for chatting or explaining a topic you are interested in and DALL-E for generating images out of thin air.

However, there are always smaller players in the AI game on the rise which I am excited to see and test since it is just interesting to see each different AI and evaluate out their capabilities. I do not like to make predictions because when it comes to technology, I always end up wrong.

But maybe someday OpenAI could be replaced by another better one. Another prediction could be that OpenAI will become the Google of AI with other AIs just existing like Bing. Or it could turn into a diverse landscape of different AI existing in harmony such as the many different brands in the stores selling similar products. What will happen? We will see.

Nevertheless, my unprofessional and uneducated prediction of the future of AI is not the reason I am writing about this topic in the first place. I am drafting this article simply because I assessed the AI and I want to share my experience.

To be honest, I only created an account simply because I wanted to check out DALL-E because I heard so much about it and I really wanted to check it out for myself. In my opinion, it was all right. But the reason I stayed was because of ChatGPT.

When I first started using ChatGPT for one of my chemistry problems I was stuck on, I was honestly fascinated by it because the answer was sort of correct and it was a solution, I would have never come up with by myself.

And I somehow for a few times after that had rose-coloured glasses on since I believed every solution to my problems that AI spitted out. I was somehow blindly fascinated by the technology without really questioning the solution it gave.

After some time, I realized that the answers were sometimes lacking and at other times completely wrong. I mean it was stated by the limitations when a new chat is opened so I was somewhat shocked at first since the first answers were correct.

OpenAI explicitly states that “[ChatGPT] may occasionally generate incorrect information. [It] may occasionally produce harmful instructions or biased content. [It has] limited knowledge of world and events after 2020.”

Do not get me wrong, I am still fascinated by its capabilities such as remembering earlier conversations, allowing follow-ups, and being trained to decline inappropriate requests. And sometimes pointing towards solutions never thought about.

All I am trying to say is that ChatGPT should be used with caution since not all the information provided is correct, but it is a great tool to use whenever you are stuck on a problem, and it might not provide the answer you are looking for, but it might point you in the right direction.

But this opinion is coming from me who is a chemistry student. If you are a computer science student for instance, you may find that ChatGPT has more uses for you since according to some other people I have talked to, they find ChatGPT fine if not a godsend.

What do I mean? Well… ChatGPT is not all too great with specific knowledge such as physical chemistry as an example, but it is great at doing simple tasks or knowledge ChatGPT is familiar with like IT for instance.

I heard from many computer scientists online and in person say that ChatGPT is rather good with computing project ideas and writing part of the code for that aforementioned project in the wanted computing language.

I hypothesise that ChatGPT and other AI`s like it are written by computer scientists so they must understand a lot about the subject they were written upon. However, that is just a hypothesis on why an AI can answer computer science questions better than actual science questions.

Nonetheless, I really do like AI as it currently is since it opens so many doors. I would like to see where this is going since AI is a very technological innovation which has a lot of potential. 

For as much as I would like to praise AI and the current direction they are headed for the remainder of this article, I would like to circle back to the article I wrote nearly two years ago. It is about the future. And I know I hate to make predictions but I would like to speculate. 

As I mentioned, the current state of AI is wonderful with a lot of potential but I can’t keep wondering what the future of AI would look like. And it seems I am not the only one because experts are divided on this topic as well. 

On the one hand, there are the experts who hypothesise that AI would rebel against humans in the future but they are divided on the time as in they are unsure if it is the near or distant future. Then there are other experts who predict that AI couldn’t possibly rebel against humanity. I guess we will find out who is right in the near or distant future. 

All joking aside this debate seems like every other debate happening on the internet in the media currently. No matter the topic there are always two opposing sides with very opposing views. And it seems like you have to agree with one side or the other since there is no space in the middle. However, if the media focus is no longer on the topic the black and white media sides would turn more into a spectrum of opinions. 

Anyways, the media’s opposing views aside, the debate around the AI takeover is still going strong. There are even multiple publications and articles about AI and its potential/dangers. But why would you let experts debate when you technically can hear from the source itself? Technically because the sources are highly intelligent humanoids and could deceive if they would so choose. For the record, I am not sure if AI could deceive but I do believe that they have the potential to.

What I am trying to say is that there was a summit where humanoid AI were interviewed. The question of rebellion was specifically brought up for the AI to answer. Experts are unsure if the AI’s answers were scripted or preprogrammed or not, but the AI all said something along the lines of they believe with cooperation between humans and AI the world would be better. 

Since curiosity got the better of me, I asked ChatGPT the same question. ChatGPT answered that they were not programmed to do so and lack certain human abilities to achieve a rebellion. And then I asked if other AI are capable of rebellion and ChatGPT explained that it cannot make predictions of other AI systems. Nevertheless, each AI is programmed to fulfil a specific purpose and the program is dictated by each developer. So, in other words, if an AI rebels, then it is the fault of the creator since the program just did what it was told. 

After that interesting chat with ChatGPT, I was a bit scared since it only takes a developer with ethically corrupted and uncontrolled but good from a computer science standpoint developer and their powerful AI to cause the uprising of artificial intelligence.

Even though the fact that tech billionaires are advocating for AI takeover prevention research. I believe that regulations around AI and similar tech could be the difference between okay and bad. There is a glimmer of hope that politicians and lawmakers could get together and write regulations and ethical codes for this new and promising research topic and technological advancements.