Imagine you could carry around a tiny expert in your pocket at all times that can answer any question within seconds. No matter if you need to solve a math problem, want to learn the basics of quantum mechanics, or a random question just pops up in your head. Unlike Google, this expert does not require you to find your own answer, looking through hundreds of pages to find the result you want. Instead, it delivers the best answer straight away. Sounds awesome? Then I have good news for you – ChatGPT is here. ChatGPT is an artificial intelligence that was developed by OpenAI, a research and deployment company. OpenAI is an AI research lab that has made it its mission to develop an AI that is able to effectively assist humans and improve our lives. The mission statement on their website reads, “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.” OpenAI is split into the for-profit organization OpenAI LP and its parent company OpenAI, Inc., which is a non-profit organization. The company was founded in San Francisco in 2015. The most famous of its founders is Elon Musk, but there are three other men behind it as well: Sam Altman, Ilya Sutskever, and Wojciech Zaremba.

On November 30 of this year, OpenAI released ChatGPT for users to test out for free. It is a demo version and still in the research phase, but there are plans to release a full version soon. I signed up and gave ChatGPT a try. But first, some general information.

What is ChatGPT and how does it work?

ChatGPT is a language processing tool that mimics human-like conversations. As the name suggests, ChatGPT is a chat. Its setup is very simple. When you open the chat, ChatGPI shows you examples for prompts and lists the AI’s capabilities. One capability that is pretty impressive is that the bot can remember things that were said earlier in the conversation. However, there are some limitations. For example, the AI occasionally spits out incorrect answers, biased content, and limited knowledge about recent events. [Text Wrapping Break]Using ChatGPT is fool-proof. You just type a question into the chat and wait for an answer. Depending on the complexity of the question or task you type in, it can take a few seconds for the answer to appear.

Where does ChatGPT get its information from?

The AI was fed huge amounts of information and texts from the internet, which it can now use as sources for its answers. It was trained through Reinforcement Learning from Human Feedback (RLHF). Human AI trainers provided conversations where they took both sides to teach the AI how to hold conversations that sound as human as possible. Next, the creators developed a rewards model using several model-messages as “ideal answers” and providing alternative responses, which where then ranked by AI trainers. That way, ChatGPT learned how to respond to certain questions.

What can you use it for?

ChatGPT can be essentially used for anything. Say you want to learn about Christmas traditions in other countries. Type in your question, for example, “what Christmas traditions are there in *insert country*?” The AI will then spit out an answer for you. In this case, the question and answer are pretty simple. However, you can also use it for more complex issues, such as coding or solving math problems. However, keep in mind that the answers are not always 100% correct. Especially if you are not just asking for fun but are planning to use the information for something, make sure to double check using a different source.

I tried ChatGPT out – here’s what I found.

When I tried ChatGPT for the purpose of writing this article, I started with simple questions and then moved on to more complex ones. Mind you, I am not an expert on things like coding, math, or science, so I can’t tell with 100% certainty that the answers are correct. However, I double checked using other sources, and the AI was always correct. Since I only tried a very limited amount of prompts, this cannot be generalized. In fact, it is commonly known and even stated by its developers that ChatGPT occasionally makes mistakes.

Some simple questions I asked that were easy to fact check where, “when did Christopher Columbus come to America?”, “which U.S. state has the highest population?”, and, “which country has the highest population in the world?”. ChatGPT answered all these questions correctly with no issues. Then, I moved on to a little more complex stuff such as, “can you write a simple code?” ChatGPT spit out a code that was so easy to understand that even I – who has no knowledge of code whatsoever – could check it on another website without problems. Again, ChatGPT was correct. Lastly, I came up with random questions that could be asked in exams and I even gave ChatGPT a made-up essay prompt. ChatGPT produced solid content every time.

This brings us to an issue that popped up several times during my research on ChatGPT, which is the possibility that it could be used by students to do their work for them. So far, anti-plagiarism software like Turnitin is not able to detect material copied from the AI chat. This means that students could possibly feed ChatGPT their assignment prompt and have it solve math problems, write code, or craft an essay for them. According to Forbes magazine, teachers are worried that software like ChatGPT will make cheating “easier than ever” (Forbes.com). As of now, schools do not have clear rules or protocols regarding the use of AI in place because this issue is unprecedented.

What does the release of this impressive AI mean for the future of humanity? Will these programs replace us soon, doing our jobs and leaving us with no other task than examining and approving the AI’s work? These worries are not new – the debate around AI and its implications for humans has been going on for a while. What impact will it have on journalism, for example? Will journalists stop writing their own texts and have an AI write them instead? Will our job be reduced to finding topics to report on, prompting the AI, and then editing and modifying its response? One of the assignments I gave ChatGPT during my test run was to write an article about the 2020 general election in the U.S. This is what I typed: “write a short news article about Biden winning the general election in 2020”. Here’s an excerpt of what the AI wrote: “On November 3, 2020, Joe Biden was elected as the 46th President of the United States. Biden, a former Vice President under President Barack Obama, ran on a platform of restoring American leadership, rebuilding the middle class, and tackling the climate crisis. He defeated incumbent President Donald Trump, who had sought a second term.” It wrote three more paragraphs that would have been too long to include here. I copied what the AI wrote and conducted a Google search to see whether the text was actually written by the AI or perhaps stolen from an old news article. The result: I could not find the article anywhere else, which means that most likely, the AI actually produced this text. I must say, I was impressed – but it also sparked some concerns.

Are texts written by AI better than texts written by human journalists? I have always been a passionate writer and knew from a young age that I wanted to pursue a career in journalism. Will AI become so good in a few years that both news directors and readers will prefer articles written by AI? Similar questions can be asked about the tech industry. And STEM, and so many more. What about coders? What about scientists and researchers? Only time will tell.[Text Wrapping Break]

However, I don’t want to draw too dark of a picture of the future. Scenarios like in Matrix or Terminator are highly unlikely. Humans will always be needed to perform tasks computers can’t perform. AI might become more and more “human” in its ways of communicating and might become increasingly better at simulating human emotions. However, no computer chip can replace the human brain with all its capabilities, the chemicals the body produces, and the blood that runs through our veins. The emotions that we are capable of feeling, from concern and affection to happiness and joy, can certainly be faked by an AI, but they can’t ever be truly felt. Our emotions might get in the way sometimes, preventing us from performing tasks effectively, but oftentimes, our emotions guide or positively impact the work we do. Humans conduct cancer research because they care about millions of people dying from this disease and are determined to find a cure. Artists get inspired by their own feelings and human experiences, and their goal is to make people feel something through their work. Journalists write about topics because they care about their impact and deem them important for the community they are serving.

Humans will never become entirely obsolete. AI will likely change our lives and our current jobs as we know them. Employees’ tasks and responsibilities might change, but their work will still be needed. Humans will likely remain the driving force behind even the best AIs. In order for AIs to work reliably and efficiently, humans must give them input. This could be a disease to conduct research on, an idea for a new app, a topic to write about, or so much more. AIs are unlikely to ever take initiative, invent a new app, write a code for it, do the marketing, and release the app all by itself. Humans will be needed to do market research, recognize a need, come up with ideas, and invent new programs or products. The coding might be mainly done by AIs at some point, but that does not mean that humans will be out of jobs.

In the journalism industry, we will still need the creativity and curiosity of human journalists to come up with impactful topics to report on. Even if AIs were to do the writing, human journalists will still have to edit the texts, modify them, and – in the case of television or radio reporters – present them to the audience. Because let’s be honest – who would want to listen to a computer, no matter how human it seems, reporting the evening news every night? I’m sure it will be interesting to see the first few times – and I’m not saying it won’t ever happen – but in the long run, genuine, real human emotions are what makes journalism so important. Because genuine journalism can make an impact on communities and change the world in a way that generic, computer-generated content could never do.

You can try ChatGPT here: https://chat.openai.com/chat