alinaovcharenko - Tumblr blog

[alinaovcharenko]

Time Flies

“We must recognize that it is quite impossible to keep up with everything that is gaunt on in our fast-changing world. However, we can learn a lot of very useful things about it” (Futuring)

We cannot know for sure what will happen in the future, but we can assume and prepare for it in time. This is one of the primary thoughts that stayed in my head after finishing the Trendspotting in Digital Media course. We learned to look at the modern world more broadly throughout the semester. We read books that helped us track how humanity has developed and is developing and what role technology plays in our lives and media in particular. The course helped us look at media in general from a different angle. For example, social media, which some people are still accustomed to, may very soon become our past, to be replaced by something completely new. Technologies that only yesterday seemed to be something new today already run the risk of being forgotten. This course helps you prepare for this kind of super fast-changing world. 

After the course, there is no fear or great anxiety about technology. Yes, in 10 years, our world will probably be totally different but didn't we just learn how to change a profile picture on Facebook 10 years ago, and today we are massively deleted from there? Just as some media professions died with the advancement of technology, so many have emerged. Professor Branagan helped us focus more on the positive side of technology than the negative, although he did not ignore the possible risks. 

All of the assignments we did during the course were helpful and complimented what we discussed in lectures and this blog as well. There was complete freedom to express my thoughts. I chose a bit of a complex topic for me. Still, with time I began to understand better and better what natural language processing is, how often we do not even notice this technology and how it can affect the field of advertising. 

[alinaovcharenko]

Continuing to talk about the DALL-E

DALL-E, a machine learning framework language model from Open AI, is popular now. Earlier, I told you about how brands use DALL-E in their advertising. In fact, there are many examples to date. For example, Cosmopolitan magazine took a risk and created a campaign and an entire cover using this technology. I was drawn to the inscription on this cover, "And it only took 20 seconds to take." It's not entirely true, but it sounds impressive. It's also worth remembering that today DALL-E is still in what OpenAI calls the preliminary phase.

The designers in advertising agencies should be the first to worry about as such tools evolve. We cannot say what we will have in 5 or 10 years, but already today, we can see what AI can do in seconds and at no cost to the user. To be honest, it's even a little frightening. Today, many people think about the fact that computers can take away our work. Even one such element as natural language processing plays a significant role. But what happens when we look at the big picture? I want to believe in the best, but it's essential to keep an open mind about how technology evolves so we can be prepared for anything.

P.S. And here's Ogilvy Paris "extending its 25-year ad saga for Nestle's La Laitière, inspired by famous Vermeer painting." This comment caught my attention:

This may be a slight exaggeration, but we do, indeed, now have opportunities to see and observe how a computer "thinks."

[alinaovcharenko]

“Everybody loves a chatbot, right? Wrong”

Although this article is about the metaverse, which we often mention in class, I found something interesting about natural language processing: chatbots and their future. In previous posts, I have mentioned chatbots and how they will evolve. For example, scientist Sophie Hundertmark talked about how soon chatbots would speak to other chatbots asking where to find new dresses or if we would soon have personal AI assistants in our ears.

This article takes a different approach to the future of chatbots. The author talks about how almost no one likes this technology, especially customers. People want real help, not the pretense of that help.

"We predict that chatbots will be virtually eliminated by the superior immersive experiences of the metaverse."

The article's author suggests that customer service representative avatars will provide product support for their customers' avatars. Still, it will definitely require more effort than is currently needed to keep chatbots running: "This effort could include adding sound designers to your creative teams, significantly increasing your investment in artificial intelligence, and constantly reiterating and improving your algorithm."

In my opinion, such predictions about the development of this technology have a right to exist. We can't say precisely when the metaverse will become part of our lives, but sooner or later, it probably will. We're already seeing many big brands creating entire advertising campaigns within the metaverse. Even smaller brands and ad agencies should look at not only the metaverse as a whole but also its more minor elements like customer support.

By the way, this article also mentions the increasing role of brand voice in the truest sense. The author writes that as the metaverse breaks down, brands will pay more and more attention to how they sound "to include an audio identity, complete with a preferred voice, sounds, haptics, and music."

[alinaovcharenko]

Draw Ketchup campaign

The Heinz company used the DALL-E image generator we already know and built its advertising campaign around it. They gave AI a simple prompt, 'Ketchup,' and got an image reminiscent of the famous ketchup Heinz.

Heinz's creators also asked DALL-E to make images depicting 'ketchup art', 'fuzzy ketchup', 'ketchup in space', and the 'ketchup renaissance'. In many cases, DALL-E returned images with familiar visual elements to Heinz fans. In some of the images, DALL-E's label images exactly match the profile and style of Heinz bottles. AI also drew images with splatter and flat white caps, commonly used on Heinz bottles.

I almost believed it.

And now again. It's as if Heinz used the DALL-E image generator, and built its advertising campaign on it. In fact, they exaggerated. Users on Twitter have noticed that not everything is as the company has shown. DALL-E is giving off completely ordinary images of ketchup.

I also decided to check the veracity of the ads, and this is the result I got:

Precisely the same as Twitter users. I'm not judging Heinz, on the contrary, they found an even more creative approach to advertising than just using AI (which is already creative to some extent).

But seriously, I think more and more companies will use technology soon. Perhaps it won't be limited to just creating the visual component of advertising. For example, you can already try to do research with AI. Today we are trying to find out how consumers perceive our brand. I think we will soon also include AI in audience research.

[alinaovcharenko]

Maia, an audience-building platform built on NLP

Maya (Mobile AI Audience) is one of the projects of Start.io, a mobile marketing and audience platform. The creators say this is “the industry's first audience-building platform built strictly on natural language processing and AI to address the needs of mobile marketers and simplify audience selection.” This project is an excellent example of how NLP can be used to help people in the advertising industry right now, but there are nuances.

I think programs like this can help the industry save money if they work correctly. Maya may not replace the work of marketers, but it can become an additional step in testing hypotheses or a tool to reach a new audience. Plus, it's a time saver. Yes, we can define the demographics of our audience and even make a psychological portrait of it, but there are hundreds of other segments. This is where AI can help. Also, a big plus is that Maya can search for the right audience in real time, which an ordinary person with a normal brain cannot do.

It is a pity that the price of the program is not listed on the site. It would have been interesting to know how much such tools cost. Also, I couldn't find any reviews on how real people are talking about the program. Is it as useful as the creators claim it is?

By the way, the company has access to 2.4 billion users. Several of the big transformation superbrands are in action. Now even a startup has access to such a broad audience. I don't know precisely which countries it allows you to cover, but the exact number goes far beyond the population of the United States.

[alinaovcharenko]

Any barriers?

As I mentioned earlier, AI image generators also use NLP, among other things. This article published on AdAge shows many examples of how brands are already using this technology.

But I want to focus on the barriers that may hinder advertising agencies. First, we should not forget that AI learns from existing images, and there is a risk that it will be biased. The article talks about how these platforms can produce stereotypical results. For example, "a man walking into a pet store" will turn out to be a white male. Still, to some extent, advertising can positively impact society, so creators must learn to overcome this barrier. I certainly wouldn't want to create more negative stereotypes.

The second question concerns ownership. Who owns the work? The author of the article says we need to be careful here. It's one thing if you type in "London" and AI generates an image with the typical red buses and red phone booths. But what if AI builds its new image based on the work of some modern artist? There are a lot of blind spots in this sense.

Finally, advertising agencies should not forget that AI, like any other technology, is a valuable tool in work, but it can't replace creative professionals. By the way, one of the experts made a prediction about AI image generators:

“Today, it’s largely still a novelty; by 2025 most creative teams will use them as standard practice. But if you haven’t developed fluency with these tools by 2030 you will probably be at a significant disadvantage.”

[alinaovcharenko]

AI-generated images, NLP, and advertising

Chatbots and voice assistants. These are the first things that come to mind when I think about Natural Language Processing. This AI avatar, which even has its own TikTok, shows that NLP can be much more comprehensive.

For example, in this video, he showed AI-generated images. It always seemed that NLP had nothing to do with it, but I was wrong. The first and critical step in creating such images is using this particular technology: “You write a text prompt, and the AI generates images from it. The AI understands your prompt using Natural Language Processing. And then AI creates images based on what has been trained.”

DALL·E has recently become very popular. Like GPT-3, it is a transformer language model. “We’ve trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language,” developers say.

Agencies and big brands have already begun using this technology in advertising!

For example, advertising agency TBWA/Melbourne created original posters for the Melbourne Writers Festival with DALL·E. They added passages from classic literature, including "1984" and "Moby Dick.” Moreover, the campaign appeared in OOH and digital. Also, an agency is planning to make illustrated e-books of classic novels with further AI interpretations! They did not even need the help of a designer, but only an idea. AI did the rest. 

In my opinion, this technology may soon become widely used in advertising. In the nearest future, not only big and popular agencies will utilize it, but also small ones or even independent authors. It's convenient, inexpensive, and creative. The main thing is that people do not get tired of it quickly.

[alinaovcharenko]

Let the AI do the boring work for you

It is always interesting to get the opinion of a person currently working in the industry. Mariya Yao, the author of this Forbes article, is currently a Chief Technology & Product Officer at Metamaven.

In this article, Mariya mentioned chatbots to capture leads, voice search to gain access to a broader audience, and sentiment analysis to understand customers. I have already mentioned a similar use of natural language processing in previous posts. Today I would like to dwell on two other points that the author mentioned since they are most related precisely to the future and not to what we are already using today.

AI copywriter for efficient ad generation. Mariya Yao says that AI can already create slogans and product descriptions today. This is incredibly convenient if new positions are constantly appearing on the site. In addition, she also mentions the international potential: “This tool is particularly popular among foreign companies that leverage this AI copywriter to create product descriptions in Chinese.” I can imagine that even small companies will soon use this advantage. This way, you can allow the copywriter to work on more complex and creative tasks and not monotonous descriptions.

AI writer for efficient content generation. Today, some AIs can create content that can pass the Turing test. Back in 2019, OpenAI released a huge language model called GPT-2. Based on some seed text, this model could generate realistic text in various styles, from news articles to fan fiction. And today, GPT-3 already exists. The upgraded version was created to be more robust than GPT-2 in handling more niche topics.

I started my blog by mentioning GPT-3. This technology should be paid the most attention to when predicting the future, including the future in advertising.

[alinaovcharenko]

Alexa and Ads

In September, a new Alexa feature called Customers Ask Alexa was announced at Amazon's Accelerate conference. Now Alexa will give answers not using random Internet sites but pre-prepared articles from brands. It will be a kind of advertising.

On the one hand, this is an excellent feature for brands. The customer will receive a more accurate response, and brands will be able to tell users about discounts or special offers. On the other hand… Do we want to get even more advertising than we get now? Voice assistants are still an ad-free island. But looking at the latest news that Alexa's department is losing money, we can conclude that this could be a good plan to increase the ROI of the technology.

In addition, the conference unveiled another feature called Tailored Audiences, which will allow sellers and brands on Amazon to send marketing emails directly to buyers.

Interestingly, people still rarely buy something with their voice:

“The estimated tens of millions of Alexa users around the world have historically been loathe to make purchases with their voices, with stats from one eMarketer survey showing that just 10.8% of customers used Alexa for online shopping in 2020.”

After analyzing this news and the previous ones, I can draw the following conclusion: the whole situation with voice assistants, which can potentially be a platform for advertising, is very ambiguous. Today it isn't easy to be sure how and in what direction this technology will develop.

I'm sure it's only a matter of time before there are ads in voice assistants of all brands. Will it be effective? This is another question.

[alinaovcharenko]

Siri old commercials: what conclusions can we draw today?

To track progress, I decided to look at how Siri was shown in the first commercials. Back in 2011, more than ten years ago, Apple introduced Siri, a virtual assistant.

“Siri uses a variety of advanced machine learning technologies to be able to understand your command and return a response — primarily natural language processing (NLP) and speech recognition” — Medium.

In this ad, we see different people asking Siri for help.

Do I need an umbrella in New York this weekend? Move my meeting from 3 to 4. Remind me to get milk when I leave work.

In fact, since then, the role of Siri has not changed much. After ten years, it still performs the same functions, and it cannot be said that our communication with this or another voice assistant has advanced very far. Yes, perhaps now Siri can better recognize our speech and give answers more quickly, but there is no question of any “complex” communication. In fact, ads with a similar essence can be shown today.

This ad is almost the same as the previous one, showing how Siri performs other tasks. These ads got me thinking about whether people are actually using this technology. Here's what I found:

“About 70% of iPhone users use Siri sometimes or very rarely. However, nearly all iPhone users (98%) have tried the voice assistant on the iPhone. It turns out that most people rarely use Siri because they feel uncomfortable using it when they are in public. Most users are utilizing voice assistants when they are alone at home or when they are in the car, most likely because of hands-free driving laws.”

We see that people do not use voice assistants very actively. I mentioned recently that Amazon's Alexa is also a non-profit product for the company. To date, it cannot be said that voice assistants are a huge success. Perhaps, to cover the costs, companies will soon start selling ads using voice assistants.

[alinaovcharenko]

Is Alexa reading minds a good idea?

Continuing the theme of Alexa and voice assistants…

I often see publications that voice assistants will be sentient in the future and will be able to support conversations with us and even “read our minds.” Do people really need it now? Isn't this something that can, on the contrary, repel?

This is exactly the problem we see in the Amazon Alexa Super Bowl 2022 commercial with Scarlett Johansson and Colin Jost. Interestingly, the ads show mind reading as something that people are not just not ready for yet, but may turn them away from buying. Scarlett Johansson imagines how Alexa invites her to brush her teeth, and Colin Jost does not want to go to the concert of his beloved. Even though more advanced technologies may help us solve everyday problems more effectively, ordinary people are of little interest. We all fear that our privacy may be interfered with.

The authors beat one of the people's fears and showed that Alexa is not as smart and scary as it might seem. This device will help you but not interfere in your life and reveal secrets. Your guests won't know for sure that the cake you supposedly baked all day was ordered from the supermarket.

A less innovative technology can be a more successful marketing ploy than a more sophisticated one.

I don't think Amazon just pulled this ad. It really reflects reality. People are not ready to talk and let AI interfere in their lives like that. Not in the next five years, for sure.

[alinaovcharenko]

How can chatbots help marketers and advertisers in the future?

Even though this video was posted on TedTalks back in 2020, Sophie Hundertmark predicts the future of chatbots in it that hasn't become our reality yet. How will chatbots develop?

I believe that chatbots are an excellent tool for marketers, which even today helps retain customers better and promptly provide them with all the necessary information. But in the future, with the development of technology, something more exciting awaits us than chatting with a robot that does not always understand what we want from it.

Sophie Hundertmark says that AI chatbots will talk to each other in the future: “It is not me who is asking the chatbot of my favorite online shop, but my chatbot who is talking to the other chatbots asking where my new dresses are.” This is another exciting variant of how natural language processing can be used in marketing, especially in increasing sales.

The scientist also mentions that soon we will have personal AI assistants in our ears. I wish she had indicated by what year we should expect this technology, but I am still delighted with her ideas. Personal Intelligent assistants will not only tell us what the weather is like outside but really help us with more sophisticated problems. And finally, she says that in the future, we will forget about Alexa or Google Home in the form they exist. The assistant will be literally built into the walls of our house.

Such assistants are an excellent field for marketers and advertisers. For example, if an AI assistant understands that you are running late for a bus, it may suggest calling a taxi of a particular brand or a top-rated brand.

The only question is, how long do we have to wait?

[alinaovcharenko]

Amazon's Alexa is struggling

At first glance, it may seem that this news is not related to advertising. In fact, it is very interesting that even one of the prominent leaders, home assistant Alexa, was not as successful as it seemed at first glance.

“During the first quarter of 2022, Amazon's "Worldwide Digital" unit, which includes everything from the Echo smart speakers and Alexa to Prime Video, had an operating loss of over $3 billion, internal data shows.”

Alexa has the potential to be a great tool in the hands of marketers and advertisers. In one of the first posts, I mentioned that maybe in the future, we could talk to our devices like real people. That is, artificial intelligence will be able to support us and cheer us up, including advertising something to us in the way our friends do now when they recommend a new TV show.

And what now, when we see that even the market leader is losing his money?

"Alexa is a colossal failure of imagination," one former employee said. "It was a wasted opportunity."

Under this tweet, people share how they use this device. Someone asks about the weather; someone turns the lights off and on with the help of Alexa. Many people turn on or off the music or use it as an alarm clock. People in the comments use Alexa quite actively, but the device is still not as successful as we would like. I've never used Alexa, but I have Siri. I still rarely use this feature, although I probably should.

Perhaps people don't really like interacting with a robot that can't react emotionally or doesn't always understand what you want from it. Maybe it still takes time for such technologies to become as much a part of our lives as a smartphone.

[alinaovcharenko]

Different ways of implementing NLP

The more I research this topic, the more I notice how natural language processing is already being used. Yes, it may not be incredible (Sci-Fi movie kind of incredible) at first glance, but this technology is already removing a lot of tasks from people or will do it soon.

For example, this article mentions that AI does not have to write complex advertising creative texts. This technology can help describe new products on websites, especially when your business has many of them daily. Usually, these are boring texts that are created according to one pattern. So why not outsource this task to AI? It's also an excellent tool for making the right keywords for your blog content. We know how important it is for your blog to appear in search results and gain popularity.

In addition, NLP systems such as mobile voice analytics can recognize patterns and tones in a person's voice and present information to them accordingly. I don't know if real businesses use this technology, but it sounds pretty exciting and promising. For example, sometimes it is clear from our voice that we want clear information, and sometimes we are ready for more "relaxed" content.

Bots. Many businesses are already using bots to speed up their business processes. I think everyone got an automatic reply to their message. But it seems we will soon meet bots speaking to us in an increasingly less robotic voice and more and more like a conversation with a real call center assistant.

In general, it is interesting how one technology can help solve and improve a variety of tasks. And this tweet confirms it!

[alinaovcharenko]

Personalized ads

This article on Medium gave me a new idea about how NLP can be used in advertising. Previously, I posted that in the more distant future, we might talk to AI as our friend and how today this technology is already being used by tools such as Social Studio to determine the user's sentiment towards different brands or topics.

This phrase in the article caught my attention: “It be used to create personalized ads based on what you are browsing online!”

It is another way how this technology can be used in the field of advertising. AI will not only be able to analyze what we are looking for and what we say about the brand, but it will also be able to instantly create the correct text at the moments of the highest receptivity. That is, we will not need to wait until the copywriter comes up with an idea and create the tagline. Perhaps AI will create here and now not just a selling text for the big target audience but specifically for me. For example, I will be able to see ads that make sentences the way I do or that implement "my" phrases or slang.

On the other hand, this technology has its dangers. How can we control everything that AI can produce? Sometimes people, even after numerous and multi-level checks, make mistakes that cost the company's reputation or can do much harm to it. A lot of people, even big brands, have experienced this. So will AI be able to sense which advertising is acceptable and which is not? What can push a person to buy a product, and what can cause unpleasant emotions or hurt their feelings? It's walking on thin ice.

I would consider this idea as one of the potential use cases for the technology, but still, we should not forget about the risks that it can carry.

[alinaovcharenko]

More than a machine, a partner

This advertisement, aka mini-movie about Toyota Concept I, is about one man and the role of an intelligent car in his life. The video creators themselves indicate that the movie is intended to depict the future of vehicles, and technologies in this movie are based on current research and development efforts at Toyota Motor Corporation. It should be noted that the film was released back in 2017, so since then, we have moved even further.

In advertising, we see how the car becomes like a close friend to a person. The car witnesses the birth of love, family trips, and quarrels, and in moments of joy, sadness or tension, it lends a helping hand. The machine supports a man with a word or turns on the right song that he didn’t even think about it. It saves him from a quarrel or makes jokes. Looks fantastic. I still have not figured it out myself, and I don’t know if it’s good that we may have honest dialogues with “voice assistants” if this will make our society more isolated, but despite everything, it looks like magic.

Turn on the light with a spell? Meh. When artificial intelligence can hold a real conversation… That sounds cool.

It seems that technology will become closer to us in every sense. Maybe soon we can have honest conversations with our phone or car. Not "Hey Siri, show me the weather," but "Siri, are you there? Something is so lousy to me, let's talk." And perhaps, Siri or Alexa will start a conversation with us themselves.

I see great potential in this for marketers and the advertising industry. We will be able use these technologies to advertise brands and get to know the audience better. Of course, we probably will face many problems with regulation, laws, and data security. At the very least, it's pretty interesting to watch this process.

[alinaovcharenko]

Have I already used NLP in advertising?

The article describes some of the exciting benefits of natural language processing and the impact of this technology on marketing and advertising. This technology can directly affect how clients communicate with the business and internal processes. For example, using NLP, you can analyze the reviews that business customers leave on social networks. Of course, this is much easier to do, not manually, but with the help of AI.

I want to note that today AI can find specific words and analyze moods and feelings. In addition, marketers can get to know their audience better. What worries people today, what do they think about, and what problems can we solve? For example, an audience research company like GWI or Simmons can give us the big picture but won't tell us what our audience is thinking now.

This article showed that I could sometimes not even think about the fact that I am already using AI, even during the educational process. For one of the assignments, we used Social Studio, a tool that helps analyze what people say about a brand. The program finds the right mentions in social networks and determines the sentiment and people's attitudes toward the brand.

“You can train classifiers in Social Studio Automate to use natural language processing to isolate relevant social media posts that contain ambiguous content.”

On the one hand, these technologies are already being used, for example, in programs such as Social Studio. On the other hand, we can look further and suggest that perhaps in the future, we will be able to recognize the mood not only by the text but also by the way a person speaks.