AI image-generation and artwork has been one of the fastest, most interesting emerging technologies that have taken over the world lately. I discovered it through my daily dose of TikTok, where people post their creations of self-portrait artwork, animals in fancy suits, and astronauts taking selfies. I was shocked with the level of detail of these photos and how real they looked. I was even more shocked when I realized that a computer made these photos.
The company OpenAI founded by Elon Musk created an impressive software called Dall-E that is able to create custom images based on prompts inputted by the user. For example, a user could input “cat wearing a bikini dancing on the beach”, and it would create a custom image of a cat wearing a bikini dancing on the beach. Users can input almost anything and Dall-E will work to make it happen.
As for the science behind Dall-E, it uses technology similar to those used in voice recognition systems and more:
“DALL-E is what artificial intelligence researchers call a neural network, which is a mathematical system loosely modeled on the network of neurons in the brain. That is the same technology that recognizes the commands spoken into smartphones and identifies the presence of pedestrians as self-driving cars navigate city streets,” explained Cade Metz for the New York Times.
Dall-E, which was launched in April of 2022, kick-started an AI-artwork movement that led some of the world’s top tech companies to start their own projects in the sector. In September of last year, Meta (parent company of Facebook, Instagram, and Whatsapp) created Make-A-Video using the same concept as Dall-E, but used user’s prompts to create short AI-generated videos.
CEO of Meta, Mark Zuckerberg explained how Make-A-Video presents a whole new ball game to the AI-generation movement by using new complex technological processes:
“It’s much harder to generate video than photos because beyond correctly generating each pixel, the system also has to predict how they’ll change over time, Make-A-Video solves this by adding a layer of unsupervised learning that enables the system to understand motion in the physical world and apply it to traditional text-to-image generation,” said Zuckerberg on Facebook (pcmag.com).
One AI program that recently took social media users by storm is called Lensa, a mobile app that generates portraits based on photos users submit of their faces. Users can buy 50 custom portrait-art images of themselves for $3.99 and even more images for a variation of costs. The artwork is extremely detailed, realistic, and overall intriguing.
Users posted the results all over social media including TikTok and Facebook, leading Lensa to become one of the most popular artificial intelligence artwork softwares of 2022. It really represented how far artificial intelligence is coming along and is only just starting out with regular use within society.
Although the app is fun to play around with and post results on social media, some concerns have been voiced about how the app could produce sexualized images of women who decide to use it.
“Olivia Snow, a research fellow at UCLA’s center for critical internet inquiry and professional dominatrix, told CNN the app returned nude images in her likeness even when she submitted pictures of herself as a child, an experience she documented for WIRED,” reported Zoe Sottile for CNN.
Other concerns have risen within communities of color and body positive advocates dealing with how the application whitens their skin tones and makes their bodies thinner, leading users to be weary of the representation and biases within the app.
Like we discussed in class, there has lately been controversy surrounding the use of artificial intelligence and where to draw the line between what is considered “cool”, and what is considered “creepy”. It seems to be a danger to artists everywhere who rely on their creativity to get by, and with a computer being able to create anything it is prompted to do flawlessly within minutes, it takes away the hard work and originality and artists cannot compete with the low prices these programs are offering. Not only that, but intellectual property and other factors such as producing harmful content and misinformation have led people to push for more governance and rules regarding platforms such as Dall-E.
“Just because images are created by AI doesn’t mean it shouldn’t be governed by ethical principles or abide by national laws and regulations e.g. privacy laws, child protection laws, copyrights,” said Hubert Halopé, the World Economic Forum’s lead platform curator for AI and machine learning (weform.org).
Although we still have a long way to go with AI artwork and image and video generation programs, where we are now surely has impressed the masses by turning their random, creative thoughts into reality. As this type of technology progresses, I wonder how the government will respond and what new laws and regulations will be created in order to keep it under control.
Hey Angela! Your blog this week was great. I loved the AI generated image of the cat playing chess. The rise of AI in artwork and pictures is so cool and I have seen it all over Tiktok.
Very interesting. I would say that I actually do not like the accessibility of this AI tool. I believe that the benefits of scientific or artistic use are invaluable, however, people may use this tool to cause disruption to people’s lives. For example, I believe this could potentially be harmful to peoples’ careers as false imagery could be developed to tarnish one’s name. I do believe this falls in the “creepy” category as images can be disturbingly altered at such an accessible rate.
Nice post. Dall_E was definitely the topic of a lot of conversation last semester. It was the Chat GPT of Fall 2022. It’s been really impressive to see the advances that we’ve made in AI in recent years, and I think MUCH more is ahead for the future. Nevertheless, I do think there are areas that AI can’t do, and we’ll be talking about that next week.
Hey Angela this was a great topic to talk about for your blog! Over the Tiktok craze of AI images and portraits came a very weird experience for me which you talked about in the later part of your blog. I came across the AI portrait which took my face and made it into some scenic portrait of myself, but for some reason kept sexualizing even the most harmless pictures of my and my friend’s faces (both boys and girls). This was very odd to me as it didn’t have the concept of what is acceptable and not which I found quite disturbing. I did hear about the biases in the AI with skin color or body image but now seeing this it makes sense due to what was happening with my experience. I wonder if regulation in these, especially on TikTok which children can use, will happen sooner rather than later.
I am curious how the government will come to regulate these programs. There is a freedom of speech aspect however there is the complexity of trademarks. I think this has a chance to be educational but that would probably be the extent of the technology and could end up bringing more harm than good as the technology continues to develop.
Hey Angela, this was a very interesting topic to blog about this week. I enjoyed reading your blog post and started to wonder how and with what are they training programs such as Dall_E? I understand that AI programs need a lot of images in order to be properly trained to handle any type of input that the user is trying to request. Aside from that, I also considered the legal implications of the art they are using to train these AI programs. One can assume that some of these training images are copywritten to protect the work of the artist. How will this affect the quality of programs like Dall_E and how could they also protect the likeness of generated work to copywritten work. This is an intriguing topic because it seems like some regulation is forcoming to programs like this in order to protect and differentiate artists from AI-generated images.
I used DALL-E the other day to make an image of a baby chicken wearing a cowboy hat singing karaoke and it was adorable. I do think this is a technology that can go both ways with pros and cons. On the one hand, it is very fun to play around with and can be really helpful for people who are more left-brained and have a harder time visualizing their artistic ideas. On the other hand, I agree with the other comments that the issues with AI sexualizing women, having race and body type bias, and just being plain creepy will have an impact on the future development of AI imagery.
Hi Angela, this is a very interesting topic. I’ve been seeing different examples of AI artwork on TikTok and noticed they are often times sexualized for some reason. I am curious to see the regulations that will be put into place in the near future.
Hey Angela! Great post!! I learned about the Dall-E software this summer and loved learning more about it through your blog. I never considered the implications of how the software could be used in a negative manner, but that is definitely a valid concern and something to consider for these open-source AI companies. Very interesting topics covered in your post and gives me a new perspective to think about!
Hey Angela! This is a very interesting topic to me. It’s honestly crazy to see that AI has been recently “popping off”. Also, it’s crazy to me how AI artwork and image programs are so accurate. It’s scary to think that in no time, AI artwork programs will take over the jobs of graphic designers; because who would pay a lot for a design that will take a while when an AI can generate it for free/low-cost in a couple of seconds?