AI has been in the news for a while now. Everyone talks about it, of course. 🤖But who feels truly up to date with these new techniques, which are constantly multiplying and improving?
A real revolution or a simple fashion phenomenon? Extraordinary potential or danger for humanity? Communication tool or generator of fake news? 🤔 The Club We Are COM went to explore the new tools of artificial intelligence for you. ChatGPT, Midjourney and company… what to expect? 🔮 Answer here!
✅ To go further in AI best practices, do not hesitate to consult our instructions dedicated to the Art of the prompt on ChatGPT and Midjourney.
AI, shall we rewind?
Finally, the origins of AI are quite recent. The first attestation of the term “artificial intelligence” dates back to physicist Alan Turing, who in 1950 established the “Turing Test”. A test which concluded that if 30% of humans participating in a conversation could not differentiate between a man and a machine, then the machine won the test.
The phenomenon seemed to have started and the promises of these new technologies were at the heart of all the issues. However, from 1965 onwards, artificial intelligence was put aside. It was not until the end of the 1990s and the creation of the Deep Blue software that AI became talked about again.
Since then, AI programs have been opened up to the general public: Google Translate, Siri, etc. The year 2014 saw the creation of the first chatbot having managed to successfully pass the Turing test. Now, more than ever, artificial intelligence is improving at great speed.
Today, artificial intelligence news is daily. We have reached a technological point and computing power that allows us to move into another dimension.
AI, the notions
According to the Council of Europe, artificial intelligence corresponds to “the set of sciences, theories and techniques whose aim is to reproduce by a machine the cognitive capacities of a human being”. In other words, it allows machines to produce and connect texts, images, videos or all other kinds of content, like human consciousness. However, AI is not yet capable of making sense.
Artificial intelligence is an integral part of our daily lives: computers, telephones, cars... and even Social Media. It is present in all algorithmic and therefore autonomous driving technologies.
Today there are 2 types of AI, distinguished by the way they record information. So, machine learning or deep learning?
☝️ In machine learning, “machine learning corresponds to algorithms, which adjust the parameters of their calculations according to the examples given to them, which allows them to adapt their operation to the data provided. » (Definition from the CNIL – National Commission for Information Technology and Liberties). Machine learning therefore allows an AI to learn based on the information transmitted to it. This is notably the technology that Amazon uses to recommend targeted products to its consumers.
✌️ While deep learning is a technology that allows artificial intelligence to learn on its own. According to the CNIL, it is “an automatic learning process using artificial neural networks and having several layers of neurons. These algorithms have very many parameters, they require a very large number of data in order to be trained. » Deep learning has the ability to analyze elements on a large scale and classify it, like ChatGPT. It is thanks to this analytical capacity that an AI can now differentiate a cat from a dog.
Is this technology revolutionary? Not for everybody. Yan Le Cun, artificial intelligence researcher who won the Nobel Turing Prize and currently Head of AI at Meta, believes that deep learning is not revolutionary. According to him, this technique only reinvents things, without yet being able to create them. In other words, it obeys probabilities but does not yet know how to make sense.
An example of deep learning: how does ChatGPT work?
AI, the market
AI, the actors
It would be impossible to list all the AI players in a market that is evolving day by day. However, remember that there are currently 2 major pure technology players: OpenIA and Midjourney.
Originally, OpenIA was an association, founded in 2015 by AI kingpin, Sam Altman and the very famous Elon Musk, who left the adventure in 2018, judging the transformation of the association too slow. However, over the past 2 years, OpenIA has accelerated, developed and launched two the software solutions : ChatGPT for content textual content and Dall-E 2 for visual content.
The research laboratory midjourney, launched its first public product of the same name, in 2022. The use of this artificial intelligence tool, dedicated to the creation of visual content, is more technical and therefore a little less widespread than its textual counterpart ChatGPT.
In addition to these two big names in AI, many companies have also launched by integrating artificial intelligence into their software. This is particularly the case of Microsoft and its search, Bing. By partnering with OpenIA, the IT multinational is changing the way we search for information. From now on, theInternet users no longer simply awaits the result of a query, but a real response, such as the summary of a feature article or an explanatory visual.
This massive partnership will allow Bing to improve its image. Moreover, the search engine goes further in this direction, since it now allows those who wish to use the latest version of ChatGPT for free.
Microsoft also plans to integrate AI into the office suite, with Microsoft 365 Copilot. It will therefore be possible to make queries on PPT or WORD, in order to generate, summarize and integrate all kinds of content. A technology that will revolutionize our daily office automation.
Meta is often compared to the metaverse, however over the past 6 months, the company has turned more towards AI. Meta has decided to launch its own “language”, that is to say its own artificial intelligence base, which aims to be faster, more flexible and less energy-consuming than those of competitors. How ? By integrating AI directly into their solutions: their social networks and their advertising tools.
Despite some delay in this area, Google ended up launching, in turn, by integrating AI into its search engine and its Workspace suite. Regarding the latter, Google's solution seems extremely interesting and promising for communicators. We can't wait to see what it turns out! 😊
And speaking of tools for communicators, Canva also now offers modules integrating artificial intelligence: creation, translation, synchronization... many new features, still non-existent a few months ago, have appeared on the platform.
Who says Canva, says Adobe. This major competitor in content creation has also positioned itself on AI. Adobe offers to generate visuals based on a library of completely royalty-free images, and therefore themselves, royalty-free. In other words, on Adobe Fly, still in the testing phase, the illustrations you produce can subsequently be used for commercial purposes.
AI, brands take action
#1- Bescherelle
Bescherelle reminds us with humor that grammar and spelling remain essential to the art of prompting. The best part of this campaign? All the crazy visuals were actually generated on Midjourney.
#2 – Mint Mobile
Ryan Reynolds is a fan of “not very funny” jokes, as we know. But what we didn't know is that by asking ChatGPT to write him an advertising text, at the time when he was CEO of Mint Mobile, the AI managed to surprise him by generating a “not very funny” joke itself.
#3 – Heinz
For AI, which is inspired by a corpus imbued with the collective imagination, a bottle of ketchup is necessarily a bottle of Heinz ketchup. Immediately, the brand seized this opportunity to position itself as a reference and decided to invite all its communities to create their own stylized ketchup visuals.
#4 – Never Without Them
The corpus of artificial intelligence can also be a barrier to diversity. This is what the Jamais Sans Elles association demonstrated, on the occasion of Women's Day on March 8. Indeed, if you ask Dall-E 2 or Midjourney for a CEO image, you will have a 90% chance of generating images of men. And vice versa for a secretary request.
#5 – Nestle
For its yogurt brand, La Laitière, whose mascot is taken from the famous Vermeer painting, Nestlé decided to use AI to recreate the imaginary behind the scenes of the painting. A creative and friendly hack.
#6 – Undiz
The Undiz brand was the first to achieve a real artificial intelligence campaign. The idea? Using AI to generate beautiful underwater visuals, a fairly complex task with pure photography.
However, AI often has difficulty generating naked bodies. The challenge was therefore significant and the outcome almost plausible. Obviously, the brand has indicated the use of artificial intelligence on its campaigns display.
#7 – Renault
The Renault brand, for its part, encourages the great public insurance to use AI tools by putting it at the forefront challenge of “reinventing the Twingo”.
#8 – The Red Cross
Conversely, The Red Cross does NOT use AI to denounce very real poverty. un ingenious way to attract attention and raise awareness.
#9 – Amnesty International
Amnesty International, on the other hand, has created a bad buzz with its campaign of photographs generated by artificial intelligence. The public quickly revolted. Why use AI to denounce existing violence? What's the point of making reporters around the world take risks? The NGO finally had to apologize and delete all the visuals from this campaign.
#10 – The FIRST 100% AI advertising
And finally, did you see the first advertising film for an American beer, entirely generated by artificial intelligence, without any actor or director ? Is this what the future of advertising will look like? We grant you, for the moment it’s still a little scary… 😊
AI, the limits?
Detecting the true from the false
The limits of content production
Certain rules still limit the production of content: violence, nudity... everything that falls under censorship. For example, you can't ask an AI to represent the Chinese President. But what will happen tomorrow?
Do artificial intelligences violate intellectual property rights?
Indeed, the corpora being created by humans, the question is valid. Generally speaking, it depends on the legislation of each country. However, certain complaints have already been made and are currently being studied by the CNIL.
Does ChatGPT violate the legalization of personal data?
Here again, the subject is debated. It is difficult for legislation to follow and adapt to new technological tools whose uses are constantly evolving.
Will artificial intelligence replace artists?
The artists themselves do not agree on the subject. For some, AI threatens the artistic spirit, for others AI has no creative force, it only knows how to merge things that already exist.
Will artificial intelligence replace communicators?
Not yet and fortunately! 😊 The Accor group put real creatives in competition against artificial intelligence. And humans won.
Will artificial intelligence replace humans?
For the moment, it is difficult to grasp the potential of the AI of the future. Artificial intelligence will not replace humans, but will certainly change the way they work. Ultimately, the skills of tomorrow will perhaps be the art of prompting, proofreading and knowing how to talk to robots.
AI fun facts
AI recommendations
