We don't have the technology yet Metaverse, as the tribulations of Meta show. Just as we don't have it for the self-driving car, as the frequent Tesla Autopilot glitches show. However, someone is working hard to bring these technologies out of the laboratories and perhaps we will soon see something happen in our experience too. It's about technologies important, transformative and general that require their own good development time.
In the meantime, though, we can enjoy some AI consolation prizes. These are hybrid technologies or subordinates of the still imperfect majors. Among these there are some that elicit a "vow", such as theGenerative AI which is the new wave in the ocean of artificial intelligence.
Tear a "vow" also because it is an open technology and can increase the creative level of our productions and bring benefits to everyone and not just to the elite of thehigh tech.
writes Kevin Rose, the technology columnist for The New York Times, “In Silicon Valley, crypto and the metaverse are out. Generative AI is in.”
The creative side of AI
Generative AI is an “umbrella” term that encompasses a branch of artificial intelligence, let's say the creative side. Generative AI can autonomously produce texts, images, videos and code starting from simple inputs in natural language, i.e. in English. For now the aspect that works best is that of image creation.
Here is an example of textual description of an image that we want DALL-E 2 (pronounced Dalì), one of the most popular applications, to produce. In a text box we can describe what we would like to achieve, for example "A bowl of soup that is a portal to another dimension as a digital art". Having received this input, the system produces a hyper-realistic image like this. DALL-E has a database of 650 million images where to take the subjects and then merge them with a technique called diffusion model.
Amazing, isn't it? No wonder this technology has become the vanguard of AI.
An instant hit
2022 really seems to be the decisive year for generative AI applications. Unlike the NFT and of the metaverse which are still an elite phenomenon, these applications have impressive numbers of users.
DALL-E 2, the image generator that OpenAI (a laboratory linked to Microsoft with Elon Musk among its founders) made available in January 2021, already has nearly two million users creating more than three million images a day.
midjourney, another image-generation application (more versed in artistic creation), was unveiled in May 2022 and today claims to already have more than four million users flocking to its Discord server.
Google, Meta, and Microsoft have also built their own image generators, but haven't released them yet. They are called respectively Imagen, Make-A-Scene e NUWA-Infinity.
Generative AI had become popular several years ago with the release of even complex text generation tools like the GPT-3 software, which gave a Google engineer, who was using it in depth, the sensation of being faced with a sentient "thing".
In a post on Wired, futurist Kevin Kelly talks about over 20 million images produced a day with generative AI applications such as Midjourney, Stable Diffusion, Artbreeder and DALL-E and comments: "With a brush in hand, artificial intelligence has become a wow engine.
A new FoMO
These numbers are a delight for the palates of venture investors always in pursuit of the well-known FoMO (Fear of Missing Out) which often turns out to be a "capital grinder".
Stability, reports Roose, la startup which just a few months ago presented Stable diffusion, an image generation algorithm, has recently raised over 100 million in funding with a valuation that already makes it a unicorn.
Always Roose reports that Jasper, an application for the copywriting aimed at marketers, has already raised $125 million and achieved a $1,5 billion valuation.
Other start-ups have scraped together massive resources for applications in industries like gaming, programming and advertising. Sequoia Capital, a major venture capital firm, wrote in a blog post that it believes generative AI could generate “trillions of dollars in economic value.”
The case of stable diffusion
To date no generative AI project has created as much hype as stable diffusion.
This is primarily due to the fact that, unlike many similar projects carefully supervised by their creators, Stable Diffusion is a open source project of free use. This means that anyone can access the source code, can download it and develop a modified version. According to the company, more than 200.000 people have gotten hold of the code since it was released this summer, and millions of images have been created using tools based on the Stable Diffusion algorithm.
And so far so good. However, there is one aspect that is creating quite a few problems Stability. These are problems we know as well as public domain tools are in action. Unlike other image generators, which have strict rules (via word blacklists) to prevent users from producing images with violent, pornographic, or copyright-infringing content, Stable Diffusion comes with a very loose security filter, which can be easily disabled by users who build their own versions of the application.
And the production of borderline content, according to Roose's survey, is already happening massively to the point that society, worried about the consequences, has intervened by establishing a principle "do not create anything that would embarrass your mother", without however introducing effective filters.
The founder and CEO of Stability, Emad Mostaque, rejected the idea of introducing content restrictions because radical creative freedom is one of the pre-requisites for building an artificial intelligence not subjected to the domination of the big Internet companies that have built a panoptic system.
Mostaque says he believes that putting generative AI in the hands of billions of people will lead to an explosion of new opportunities. It will certainly happen.
The impact on the creative professions
For years, the widespread belief has been that artificial intelligence and automation will lead to the disappearance of the professions related to manual work repetitive. It was thought that hauliers, cashiers, warehouse workers, assembly line workers would lose their jobs to robots, while workers in creative industries such as art, entertainment and media would be left out of the employment consequences produced by the robot. TO THE.
Well, an unexpected fact has recently occurred: artificial intelligence has entered the liberal and creative professions.
Kevin Roose writes in this regard: “It is still too early to say whether this new wave of applications will end up costing artists and illustrators their jobs. What seems clear, however, is that these tools are already being used in the creative industries."
And that's right. A representation generated with DALL-E2, which consists of a fictional work that impressively blends the masterpiece of Jan Vermeer, The milk girl (Rijksmuseum, Amsterdam) with other portions of Flemish school paintings, served the agency Ogilvy Paris, which is part of WPP, the largest ADV company in the world, for the advertising campaign of the "La Laitiére" product line of the Nestlé destined for the French market.
Examples of using generative AI in the creative industry
The New York Times reporter reports five cases of creative industry professionals who are regularly using some generative AI application in their business.
Collin Waldoch, 29, a video game designer from Brooklyn uses Dream studio to create characters and items for use in online gaming. Cost: a few minutes of work plus a few cents for the use of the images. These applications have a credit billing system. For example, 30 credits (which allow you to use 30 images) cost $15.
Isabella Bears, 24, an interior designer from San Francisco uses generative AI to create mock-ups of furnished spaces. You upload the photos of the rooms that the customer intends to furnish to the Interior AI app and apply filters to obtain renderings with the new furnishings.
Patrick Claire, 40 years old, a filmmaker from Sydney in Australia uses DALL-E 2 to prepare presentations of filmic settings to submit to production companies. Find this tool vastly superior to any other tool available for creating virtual adaptations and stage design.
JasonCarmel, 49, is an executive at the Wunderman advertising agency in New York. Sometimes his team finds itself in the need to create an image from scratch, such as that of a dog at a poker table. Making it in the traditional ways would require an entire day of work for a designer, perhaps only to realize that it doesn't work. Well with DALL-E 2 it can be done in seconds. And n-versions can be made of it.
Sarah Drummond, 33, from a graphic service in London, uses Midjourney or DALL-E 2 to design black and white cartoons for signage. Previously, he spent a lot of time preparing sketches, the so-called "blob drawings". With generative AI applications it now takes him a few minutes.
… and the consequences?
In any case, all these professionals agree that generative AI is a very useful tool in creative work, but it will never completely replace the professions of this sector. At least for now, I might add.
However, there is a nice "but" to ask that arises from a number of difficult questions related to plagiarism, authorship, artistic property and the world of work. Perhaps there is one greatest of all, he writes Tom Faber in the "Financial Times", this "is the beginning of the end of human creativity?".
A good question and perhaps the best thing is not even to ask it. But one thing is certain: the history of art and creativity is inextricably intertwined with technology not only in terms of techniques but also in terms of content. And art has never been supplanted by technology.
Sources:
Kevin Roose, A Coming-Out Party for Generative AI, Silicon Valley's New Craze, The New York Times, October 21, 2022
Kevin Roose, AI-Generated Art Is Already Transforming Creative Work, The New York Times, October 21, 2022
Tom Faber, The golden age of AI-generated art is here. It's going to get weird, The Financial Times, 27 October 2022
Kevin Kelly, Picture Limitless Creativity at Your Fingertips, Wired, November 17, 2022
