【Technological Innovation World Trend】

　　2022 appears to be a breakthrough year for artificial intelligence (AI), with many innovative AI products hitting the market.

One of the most popular products is the photo-editing app Lensa, which lets users create anime-like digital portraits, a feature dubbed "magic avatars."

Not only has Lensa topped app stores around the world with these popular avatars in recent times, it's also been a hit in the art world.

　　What makes Lensa stand out

　　The app, released in 2018, offers a range of photo-touching features, such as making selfies look better by applying filters and removing "blemishes".

The latest version of the app has been updated with the new feature "Magic Avatar".

　　Users upload 10-20 facial photos through their mobile phones and pay a certain fee, and they can "entrust" Lensa to make up to 200 AI portraits with different styles.

　　Stability AI, an American startup company, open sourced an AI model called "Stable Diffusion" in August this year, which can generate corresponding images based on text given by users.

According to a report in How-to-Geek, an online technology magazine, it allows people to imitate different art styles, including comics, sci-fi, pop art and traditional portraiture.

And these portraits of Lensa are generated using "stable diffusion".

　　Australia's "Dialogue" magazine published an article saying that if "stable diffusion" is a text-to-image system, then Lensa seems to be very different because it accepts images, not words.

That's because one of Lensa's biggest innovations simplifies the process of inverting text.

　　Lensa takes user-provided photos and injects them into a "steadily diffuse" existing knowledge base, teaching the system how to "capture" the user's features in order to stylize them.

　　Users worry about privacy leaks

　　While the "magic avatar" is an interesting and impressive feature, it raises concerns about personal privacy and likeness rights.

There are concerns that the program could create facial portraits of individuals without their permission by using selfies as input to an AI model.

　　Additionally, some questioned the app's price, which requires a minimum payment of $6, or $53.99 per year, to create AI images from personal photos.

There are also concerns that users are effectively paying to train facial recognition AI and giving away private data.

　　Additionally, the app's creator, Prisma Labs, has previously been in the dust for accidentally generating pornographic images of nudity, despite the app's policy of "no nudity" and "adults only."

The company's chief executive said such behavior would only happen if the AI ​​was deliberately directed to create such content, which violates the app's terms of use.

However, as noted in a report by TechCrunch, some have expressed concerns about the app's potential for abuse and the impact it could have on users' self-image and body image.

　　In addition to the risk of involuntary pornography, there are concerns that AI could be used to create political misinformation and disrupt education.

Overall, the Lensa AI application is a reminder that AI technology is still experimental and could have unintended consequences if not properly regulated.

　　Artists worry about copyright issues

　　The popularity of the Lensa AI app and its "magic avatar" feature has also fueled concerns among artists that the widespread use of AI image generators could put them out of work.

　　According to a report by the American media Futurism, this machine learning model is trained on images without consent, and individuals or artists cannot opt ​​out of the data set.

　　However, the Dialogue article says that the images Lensa generates borrow ideas from other artists' work, but don't contain any actual snippets of their work.

The Australian Art Law Center clearly pointed out that although individual works of art are protected by copyright, the stylistic elements and ideas behind them are difficult to be included in the scope of protection.

　　Diaz, a Spanish cartoonist and illustrator, said that AI is just a very hollow remix of what already exists.

"If you look closely, there's no consciousness behind it. I really hope that people understand what we (artists) do when we create something...Hopefully people will expect human art."

　　"I'm not against artificial intelligence, if it's a tool we can use and people learn to value what we put in our work, I think it's going to be great," Diaz said.

　　Where does the future of art go

　　The "Dialogue" magazine article believes that although the AI ​​​​art model has made great progress in the past 5 years, it still needs to face many challenges.

While the words in their works are recognizable, they are often unconscious.

　　There is also the obvious constraint that these modes can only produce digital art.

AI cannot work with oil paint or chalk the way humans can.

Like vinyl making a comeback, technology may initially create a new form, but over time people always seem to return to the original form at its highest quality.

　　Ultimately, as previous research has found, AI models in their current form are more likely to become new tools for artists than digital substitutes for creative human work.

For example, start with a series of AI-generated images, which are then selected and refined by human artists.

This combines the strengths of the AI ​​art model (fast iteration and image creation) with the strengths of a human artist (vision of the artwork and overcoming the problems of the AI ​​model).

This is especially true in the case of commissioned art that requires a specific output.

　　In addition, what needs to be considered is that artists who choose not to use AI may not be able to keep up with the pace of artists blessed by AI, and be abandoned by the times.

(Science and Technology Daily intern reporter Zhang Jiaxin)