It would be apt to say that 2022 is the year of the generative AI explosion. While the upstream aspects of infrastructure, datasets, and models have been building and evolving for years, what makes this year special is the emergence of “multi-talented” models with amazing understanding and creativity, and the application interfaces built on them. These new technological creations both amaze and amaze the public, and at the same time make classic questions such as whether AI will replace human beings a hot topic of discussion again.

As the vanguard of technology tasting, we will not be absent from this AI event. Since the end of last year, the number of AI-themed shares in the community has increased significantly. From experiences and tests to insights and reflections, these shares cover a comprehensive range of popular AI products and hot topics in recent times. Here, we have selected some of the best-received articles for our readers. If you are also interested in AI, reading these shares will definitely help you gain some new knowledge, resonance or touch.

It is worth noting that in the process of experimenting with AI tools, our editors also gave full play to the spirit of critical thinking, and were not satisfied with just playing with AI tools and marveling at their potential, but reflected further on this basis, expressing their views on the relationship between AI and people, the future division of labor, and other issues.

“Artists will have their jobs; they will join forces with AI to make their work more efficient and further push the boundaries of human creativity. However, those who are engaged in non-artistic drawing creation may suffer a career crisis. In addition, I wonder what the emergence of such a powerful tool as DALLE would mean for the forgery of academic paper graphs. All I can think of is to use technology against technology. Only an equally well-informed model can defeat this kind of ‘fabrication from nothing’ that goes against academic ethics.”

Don’t expect too much from Elicit’s direct answers. We can take the limitations that Elicit finds and turn them into new questions, and let Elicit help us find others’ research in the vast research literature to determine if they are valuable questions or if they have already been beaten to the punch. This way, you can save yourself a lot of headaches. Try it out.

The purpose of knowledge creation is not for medals, but to satisfy curiosity and to make the world a little bit better. Instead of all creatures going together to compare climbing trees, “all creatures should be free”. This is the effective way to break the involution. “Laziness” is our characteristic, not a sin. If a tool can make me more efficient, I will use it to the fullest, without feeling guilty. Just be careful to “feed myself with things” and not to “serve myself with things”.

Throughout the year, voices have been raised criticizing Copilot’s ambiguous position. On the other hand, some optimists argue that users of relatively short suggested snippets are exempt from licensing under the fair use doctrine, while others are more radical, arguing that such machine-generated code is not an “intellectual creation” and therefore does not constitute a “work” for the purposes of copyright law.

Of course, as with any AI-related IP issue, there are far more questions than answers, and more white space than rules. It is still safer to use common sense, check manually, and accept with caution.

Excellent AI paintings require people to think about how to use figurative language to express abstract thinking and optimize the details, a process that cannot be separated from human intervention. Even if the AI painting technology is advanced, no profound work can be born in an empty head.

AI painting is an art of collaboration and expression. Like a computer programming language, what it can do is to help you communicate with the computer and use it to quickly show you what you want and need.

How do the rights to AI-generated works accrue? It is not that art history has not witnessed similar questions; Marcel Duchamp’s urinals and Andy Warhol’s cans were controversial when they were introduced, but ultimately did not destroy art, as was sometimes feared, but rather expanded the definition of art. One possible response is that AI users make a creative effort in writing textual cue words, thus justifying their enjoyment of authorship rights.

A more innovative formulation comes from The New Breed, which argues for a reference to the human-animal relationship in dealing with human beings and AI, for example by holding users liable for the tort consequences of AI creations in the same way that animal keepers are liable for the ‘atrocities’ of their animals.

AI cannot write our experiences and stories, nor can it write reflections that break through our genre. Whether it is the degree of innovation, individuality, or discernment, AI cannot outperform the creator; this is also my criterion for judging good content.

Although AI-generated art seems to be “imaginative” but not “precise” or “stable”, this is actually an “engineering” problem rather than the limitation of AI’s ability. If we consider painting as a means of depicting the heart and expressing oneself, AI-generated art is also a means, and AI art creators will get their own “mind-flow experience” in the process of using AI, and experience the joy and beauty of art created with AI.

Rather than seeing [AI] works as an imitation of human graffiti, it can perhaps be understood as a rethinking of human creativity – and perhaps human creativity does not come from human reasoning and subjective agency. All of this makes the proposition that “imagination is natural rather than artificial” less stable. If we think that what AI does is not new, how can we argue that what humans imagine is new, rather than a recombination of multiple existing elements?

AI composing is a challenge from the rational world to the sensual world. The fusion of AI and art is more dependent on breakthroughs in basic science, but beauty does exist, and the art created by AI today is [already] colorful. This is an excellent area for the fusion of technology and art, and a journey to unify the digital and the emotional.

The traditional “modeling-rendering” process is not as efficient or effective as it could be, and the features of AI drawing can make up for it in these areas. For us [designers], we can also make the immediate step faster by adjusting the preset style, inputting more relevant images to adjust the model, and designing tools that are more suitable for designers’ daily interaction.

AI painting service naturally has more possibilities in To B area. [Although] there are few public landing cases yet due to the limitation of model quality, copyright controversy, and the current relatively early technical stage, we believe that more To B success stories may emerge in the future in the direction of advertising industry, material library, auxiliary tools, marketing customization services, offline physical combination services, and online virtual spaces such as metaverse.

We advocate that if a living artist is specified in the generation prompt, then at least the AI-generated images should be released with CC0 protocol, and the information about the artist should be kept, and the commercial use should be kept as non-commercial as possible until authorized by the artist.