Blog

Artificial intelligence has made tremendous strides in recent years, and its capabilities have now extended to the world of content creation. From generating news articles to creating social media posts and marketing materials, AI is changing the way we create and consume content. But what does this mean for the future of content creation? Will AI eventually replace human writers and designers, or will it simply serve as a tool to enhance their work? As AI continues to evolve, it’s worth taking a closer look at its impact.

Well, the first paragraph was entirely written by the latest version of the AI language model ChatGPT. It took the system a few seconds.
The request was:

“Write the first two paragraphs of an article about AI in content creation, with a cliffhanger.”

Okay, it is not two paragraphs, but it is enjoyable, and the system even understood the concept of a cliffhanger. Not bad, right? But let’s pause here for a moment to clarify some terms.

What is AI?
You can of course formulate it more precisely, but AI is basically a piece of code which can improve without human intervention. A code writing its own code if you like. Take Google’s famous AI chess program as an example. At the development stage they didn’t feed it with the usual data of millions of chess games, they rather let it play against itself. The system played gazillions of games on its own, and drew its own conclusions. The outcome was a Chess software with a new and unique approach to the game, and it defeated everything that was thrown at it.

What is an AI language model?
A language model basically does the same. You give it tons of text, and it plays around with it. It takes out a word from the text, and then from the context tries to guess what the missing word is. If it guesses right, it learned something. ChatGPT is such a language model, developed by a non-profit company called Open AI. Open AI was founded by tech entrepreneurs like Elon Musk or Sam Altman, and since it is for the time being free of charge and accessible, it can be tested by mere mortals like you and me.

But does it have a headphone jack?
Playing around with the system instantly raises a lot of questions. Let’s look at the following example. I asked it to:

“Write a short poem about journalists”

And the feedback few seconds later was:

“Journalists, the brave and bold
Their quest for truth never grows old
They fearlessly chase the news
Uncovering secrets, exposing abuses
Their words have power, they can inspire
They can bring about change, they can set fire
To the hearts and minds of all who read
Journalists, the ones we all need”

There is a certain irony to it, a system capable of writing articles in seconds claiming in a poem that we need journalists. But beyond the future of journalists this also makes us think about more profound questions. What is art, what is creativity? According to the Oxford dictionary creativity is “the use of imagination or original ideas to create something; inventiveness.” Well, AI used an insane amount of sample text instead of “imagination”, but it did create something new, something original and inventive.

Let’s see how it can handle more complex requests:

“I have written a text in English and translated it to German. You can find it below. Does the translation need improvement, or is it okay?
The wholly grail of transcreation seems to been cracked with the use of AI. Der heilige Gral der Transkreation scheint mit dem Einsatz von KI geknackt zu sein.“

The answer 1-2 seconds later:

“It looks like the translation is generally accurate, although there are a couple of minor issues that could be improved upon.
First, the phrase “wholly grail” doesn’t quite make sense in English. The word “grail” refers to the Holy Grail, so it is already a whole or complete object. It might be more accurate to say “holy grail” or “complete grail” in English.
Second, “transcreation” is not a commonly used term in English, so it might be more clear to use a more descriptive phrase, such as “the process of adapting content for different languages and cultures.”
Overall, the translation is understandable, but a few minor revisions could make it more accurate and clear.”

ChatGPT did understand the question, it was able to interpret two separate languages involved, it managed to find the error in the original, contrary to the instruction, and offered a reasonable solution.

Is there a catch?
The idea behind the Open AI project is that such a disruptive technology shouldn’t benefit only a few. The company is providing access for the general public. Sort of. Open AI also launched Dall-E, another AI system which can create images and art from text. Tell it to create a picture of “a human supported by AI”, and within seconds you have the result that has served as an illustration for this article. It is fascinating. And by now it has a payment model. Dall-E started as a free tool for a select group of testers. If you register now you get 50 free credits, enough for 50 requests, but if you want to create more pictures, you have to purchase additional credits.
According to ChatGTP, which eagerly answers your questions on the matter, Open AI is and remains a non-profit organization, they are only collecting funds necessary to keep the systems running and the development ongoing. We have to trust ChatGPT’s word on this.

Shall we destroy the machines?
At this point all of this starts to sound a bit scary. So an instinctive human reaction emerges, deeply engraved in our pre-historic brains. Remember the industrial revolution, where workers started to destroy the machines which they thought would replace them? Well, one can say that in the end the machines won. This will probably be the same with AI. In cold heart capitalism all the actors, from states to companies strives to develop something that can offer better value at lower cost than anything before it. This is how we got mobile communication and the Flowbee. AI won’t be any different as it can do a lot of things a lot more efficiently. And don’t blame capitalism for it, authoritarian state leaders are probably also pushing hard for such solutions.

If you can’t defeat them, join them
The proliferation of AI language models is inevitable. It will have a fundamental effects on professions like journalism or even software engineering. Yes, it can write code too. And in a world where you can create tons of reasonable content in seconds, it can have fundamental effects on our society as a whole. Courts can be flooded with petitions. Democratic discourse can be shut down with a flood of counter arguments, or just slightly untrue statements. Publishing companies will move resources from content creation to fact checking as content will be abundant, but validated content will become an asset worth paying for. Society will need to build its defenses. Our politicians will have to act much faster than usual. But on the individual level the situation is not that gloomy. Instead of rejecting the tool, making use of it actually is fun.

Take the following example. If you are in the mood for a philosophical discussion, just ask AI.

“How would you extend Asimov’s laws?”

“Isaac Asimov’s laws of robotics are a set of rules designed to ensure the safe and ethical use of robots. The laws are as follows:
A robot may not injure a human being or, through inaction, allow a human being to come to harm.
A robot must obey orders given it by human beings, except where such orders would conflict with the First Law.
A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
Asimov’s laws have been widely influential in discussions about the ethical use of artificial intelligence, and have been referenced in numerous works of science fiction. However, as AI technology continues to advance, there have been discussions about the need to potentially update or extend Asimov’s laws to better address the complexities of modern AI systems. Some possible ways to extend Asimov’s laws might include:
Adding a fourth law that requires robots to respect the autonomy of humans and not interfere with their decision-making processes.
Expanding the scope of the laws to cover not just physical harm, but also emotional and psychological harm.
Incorporating principles of transparency and accountability into the laws, to ensure that the actions and decision-making processes of robots are explainable and traceable.
Developing more specific guidelines for the use of AI in different domains, such as healthcare, transportation, or military applications.
Overall, the extension of Asimov’s laws would require careful consideration of the ethical implications of AI and the ways in which it could potentially impact society.”

Not bad. My revolutionary idea on the topic originally was to add one single law, stating that machines mustn’t instruct humans. That still might do the trick with a few tweaks but I have to admit that the AI answer to the problem is much more nuanced. So you can use the system for fun.

But can it be used for work?

“Give me arguments for drone inspections of solar plants”

“There are several arguments in favor of using drones for inspections of solar plants:
Safety: Drones allow for inspections to be conducted without the need for technicians to physically climb up and inspect the panels, which can be dangerous.
Efficiency: Drones can cover large areas quickly, allowing for more thorough inspections in a shorter amount of time.
Cost: Using drones for inspections can be more cost-effective than traditional methods, as it eliminates the need for technicians to travel to the site and reduces the amount of labor required.
Precision: Drones equipped with high-resolution cameras can capture detailed images of the solar panels, allowing for more accurate assessments of their condition.
Environmental benefits: Drones do not produce emissions, making them a more environmentally-friendly option for inspections.
Overall, using drones for inspections of solar plants can help to improve safety, efficiency, cost-effectiveness, precision, and environmental sustainability.”

Drafting a sales document suddenly seems much easier.

Last but not least, let’s check if AI can help us spend our money.

“I want to buy a new soundbar for my tv. Which is the best soundbar out there right now?”

“It’s difficult to definitively say which is the best soundbar as it depends on your budget and personal preferences. Some popular options on the market currently include the Sonos Arc, Bose Soundbar 700, and the Samsung HW-Q90R. It may be helpful to read reviews online and see which one best fits your needs.”

“What is the best budget hotel near lake Como?”

“There are many budget-friendly hotels near Lake Como that may fit your needs. Here are a few options to consider:
Hotel La Perla: Located in Lecco, this hotel has a 3-star rating and offers rooms for as low as $50 per night.
Hotel San Rocco: This 3-star hotel in Bellagio has rooms starting at $70 per night.
Hotel Milano Scala: Located in the heart of Milan, this 3-star hotel has rooms starting at $80 per night.
Keep in mind that prices can vary depending on the time of year and availability. It may be helpful to compare prices and read reviews before making a decision.”

Et voilà. Tie in advertising and you can instantly offer the service free of charge to the masses. Google search, a multibillion dollar business suddenly seems very outdated, doesn’t it?

(The article was published on 17th of January 2023 on LinkedIn, well before any related announcements from Microsoft or Google.)