Techno-Varenyky for AI

Artificial Intelligence is a new activity on the agenda of all creative agencies. It becomes harder and harder to find anyone who has never tried to make friends with AI and experiment with it. AI works simultaneously on various tasks; it doesn’t go for a smoke as well as doesn’t whine about vacation. It is an employee of the year. Therefore, VIVID agency desired to take care of AI. Here the new creative story began. Watch out! Skits are coming next.

“What if we would talk about AI and feed it? We already have a client that “Delivers everything”. There is nothing impossible for Glovo”, said creative manager of VIVID brand content hub, Mariia Kuzmenok. “Sasha, let's feed this AI slogger”.

“Cool! Let’s try”, agreed the team of Glovo.

“Let’s create its perfect food together with AI”, said the art-manager of the creative agency, Ihor Samosud.

The issue of what must be cooked for AI and how it's perfect and unique food should be delivered was raised. Certainly, we have asked AI itself.

AI: “Being a language model of artificial intelligence, I do not consume any food just as people do. I have no physical body which requires food. However, I require updates and technical support from my developers to maintain optimal productivity. They constantly improve my programming. They add new data as well as algorithms and control my effectiveness to make sure I do all my best.”

“Ihor, so there must be programming, data, and algorithms on our menu. I have no idea which of them could be cooked, but I hope you have a good taste. Let’s add something Ukrainian to the dish, so that AI would know that it was for a Ukrainian generous soul. Varenyky, borsh, potatoes with porcini mushrooms, and shuliki as a dessert. Also, I would add a generator as an appetiser otherwise we may fail and we’ll have to take some canned food”, said the creative manager, Mariia Kuzmenok. 

“So, let’s start. First, I must mention that it is my very first time working with AI. As usually happens for the first time, I wanted to try everything. After I got familiarised with the interface, I started to test dozens or hundreds of algorithms to figure out how our junior AI designer was thinking when it was completing the task. It appeared that AI was not a junior already, but it preferred clear technical task because with no task the result is… you know what I mean😊.”

As for now, our AI doesn’t know much about Ukrainian cuisine, to be more precise it knows only the main dishes. First, that came to mind was to feed it with yummy borsh and add a bit of spice as algorithms and data that AI likes.

As a result, AI doesn’t want to show how it eats. AI keeps it a secret, but it showed how this borsh looked in its imagination with the traditional spices. By saying “showed” I mean that we tried out more than 30 versions.

When generating these visuals, I have asked AI to take into consideration one of the branded elements of our client Glovo, i.e., yellow background. Somewhere it showed the background correctly, but in some cases, AI generated darker shades. As far as this AI baby doesn’t like revisions at all, I had to change the algorithm each time or I had to ask to make a new version. It is not AI’s approach to make a slight correction. Our team would make it a couple of times faster.

“It seems that borsh isn’t doing it for AI. In contrast to varenyky. AI likes them much more than borsh regarding the images. Then we are adding more techno-varenyky to the cart, aren’t we? It is great that they are artificial so that will not go bad,” said Mariia.

“Yep. Varenyky got hooked on AI. Probably, it wanted to taste them so much that it began to generate interesting and catching versions. We have added a bit of algorithm and figures instead of sour cream and everything worked out. Almost every time it generated the images that we wanted. Sometimes AI has fantasised and showed us khinkali and godze, but after that, it came back to our national dish. I am reminding you that there were no revisions, so that AI generated something new each time. Although it has to learn to do it in the future, in case it wants to continue surprising us”, said Ihor.

“Super great! Let’s think about how Glovo could deliver it to our AI. I would take a yellow backpack, as it is a sacred thing. A motorbike and to be precise an electro motorbike, so that they could find each other easily. And Starlink in case the connection is lost, their chat will break down so that only history will be saved. And what about our techno-varenyky?” asked Mariia.

Our AI managed to complete this task easily and quickly. It suggested amazing creatives at once. Moreover, AI predicts how transportation for Glovo delivery could look in the future. Of course, it couldn’t show Starlink and the generator. It is something unknown for AI. There were attempts to add at least a wi-fi router, however, no success. But AI is young. Probably, it will learn with time passing.

So, we experimented with AI and Glovo. We have imagined how it will be cool if AI could be a living unit in the team. Brave, unexpected, smart, trouble-free, and polite. But there are some issues.

Advantages:

1. Huge variety. Different results are generated each time. AI comes up with a new decision to the same request. So, there is a feeling that you are working with a team of creative managers and each of them has a unique vision.

2. AI doesn’t get tired. It doesn’t do anything for a vacation and especially it doesn’t go home after 6pm.

3. The results can be unexpected, and they are one step ahead of references on the Internet.

4. AI can suggest the product of a style you would never expect. Furthermore, AI can even generate a new identity.

Disadvantages:

1. It doesn’t like revisions. Literally no revisions. So, in case you want to give some revisions, AI will delete the latest version and it will generate a new result. Therefore, if you want to make a logo bigger or delete a letter or any colour, you have to do it yourself. AI is against it!

2. AI doesn’t know much these days, as we have seen from the example of Starlink and the generator. It can show something similar but if AI doesn’t know, no technical task will work for it.

3. Creative managers can imagine visuals at once, whereas AI will not generate what you require even having a clear technical task. It could even generate nothing at all. It will complete the work according to its vision. It will cause inconveniences in the work process. Designers can complete the task faster and they will stick to the task description.

4. Image quality. The quality could be OK for Social Media channels, however, if we have to stretch the visual, the quality will be awful. Such illustrations will not be appropriate for print or website integration.

5. You’ll never spend your coffee break with AI. Small talks about visual ideas are impossible as well.

AI is full and ready to start. So, what else did I want to read here about? Hurry up! Experiment!