I Used Artificial Intelligence to Write an Article and Here’s What Happened.
Hey, artificial intelligence seems to be all over today’s world. These fake, silicon brains have been replacing our ‘feeling brain’ slowly but surely. Taking on more and more tasks off of us. But, have you ever wondered about the limits of A.I. What it can and can’t do? That’s the point where I decided that I’m going to run an experiment. The purpose was to find how good modern machine learning really is. So, I decided to try to get an article written exclusively by a computer.
The Experiment
The first step was to research my options. One option was to try and build the A.I. on my own, and the other was to find a service online. I did some research on how to create an article-generator from scratch and it turned out that it wasn’t that hard to create one, but surprisingly hard to train one to actually be good. I finally decided to go with a ready-made alternative that I found on Google.
The service was easy to navigate and work your way through. That’s when I started realizing just how much time I had saved, had I worked on coding my own. It was so easy, that I was almost surprised when I first used it. You just supplied the topic or niche that you want your article to be about, and that’s it. There was also an advanced option that let you insert a paragraph, which the A.I. would use as a guideline to write on.
I decided that I would use the advanced setting, since there was no ‘productivity’ in the topics, and that was what I wanted the A.I. to write about. So, I wrote a quick paragraph about productivity and mindfulness, and the journey began. I waited about 20 minutes for the algorithm to generate my article and, needless to say, I was quite nervous about how it would turn out.
After a small coffee break, the wait time was finally over and I was ready to see what was supposed to be a full article. I tried reading the first paragraph and it didn’t really make sense. I continued on reading and, to be honest, nothing made sense. Nevertheless, I published the article on Medium just like I had promised to do, and now, you can actually find it on my profile under the title ‘How I use guided meditation to be more productive. The article was worthless, nothing made sense, the paragraphs were all messed up, and I was disappointed.
But, that’s not all, I was determined to make this work. I started fiddling around with all the options that the service had to offer. I tried writing business articles, health articles, and even spirituality articles. Somehow, they all turned out moderately good, a miracle compared to the first one. Now, you’re probably wondering, why didn’t you post the better ones? However, the original challenge was to publish whatever article the A.I. gave me, and since that was the first one, I knew I had to do it, and you had to see it.
As if A.I. writing articles wasn’t interesting enough, I wanted more. I wanted to learn how machine learning operated, how its internals worked, and why it was so inconsistent. That’s when I realized that the huge disparities in the language of the computer were due to the guidelines that I had provided. Whenever I entered a paragraph to be used as a platform to build on, the algorithm went crazy. When I let it do its thing, however, it wrote articles that were at the very least readable.
The Future of A.I.
“[AI] is going to change the world more than anything in the history of mankind. More than electricity.” — AI oracle and venture capitalist Dr. Kai-Fu Lee, 2018
Now, you’re probably wondering to yourself, what was so important about this experiment. The answer is very simple, testing the limits of brand-new technology is always useful. This experience helped me understand that machine learning has a lot more to learn and accomplish. It’s still in its early phases of development and we should be expecting much more from it.
A.I. seems to be everywhere, from your fridge to IBM’s supercomputers. Still, it has a lot of ground to cover. Once that happens, researchers think that we will see a new revolution, where A.I. takes over millions of jobs, but just like past revolutions, millions more are created, in industries that we didn’t expect.
In the next few years, the scene of Machine Learning will probably change drastically. We will probably see technologies like unsupervised learning get developed, changing the way and rate at which computers learn and the way we interact with them. Engineers will not have to input data manually, instead, automated algorithms will gather data on their own, regardless of the interaction we have with them.
Something important is that we now see articles titled “The AI-Enabled Future”. Those offer an optimistic view on the topic of Artificial Intelligence, and I think that that’s the correct way to think about it. Most likely, we will not see a Terminator-like dystopian world where AI takes over and the human species goes extinct. The most likely scenario is a future where A.I. contributes to the world and helps us solve critical problems that humanity will have to face in the future.
The Aftermath
To conclude, artificial intelligence will make leaps forward, but I would like to raise some concerns about the future of article-generating algorithms. One of them would be if the A.I. infringes some kind of copyright, should we blame the author or the system? Also, if the computer makes an offensive statement, like using racial slurs, is the author to blame here?
