Artificially-intelligent image generation has become a hot topic in recent weeks and months. Not too long ago, an artist created a stir for winning a competition with an image that he largely created using an artificially-intelligent tool. Now, Microsoft has announced that similar AI-powered image generation tools will be integrated into many of their applications.
Ask any designer, writer, or other content creator you know…the blank page can be daunting to face down. The intention behind these new tools is to help give a user something to put on that page that they can then use as a springboard.
Take the new Microsoft Designer application that is being released—AI is functionally built into the software, with users having the capability to provide a description of what they need to the DALL-E 2 AI. The image generation platform will then provide numerous generated images for the user to select from. Designer was originally a feature within Microsoft PowerPoint, but its utility proved sufficient for a standalone application to be developed.
Microsoft also has plans to add Designer into Edge and Bing to allow users to access these capabilities as they create social media posts and custom images.
There are assorted concerns that this kind of technology presents that Microsoft has needed to consider as they have developed these features—both in their existing software titles and in those newly-developed ones, like Designer.
First off, AI-generated images as a practical concept are not without some controversies. In simple terms, the platforms (referred to as image synthesis models) used to create these images do so by effectively compiling the millions of images on the Internet that relate to the prompts…the vast majority of which are protected by copyright laws, and it isn’t as though your AI platform is securing permission to reference each and every image from each and every respective owner.
It’s almost as if you went online and looked at Edvard Munch's famous painting The Scream, and then immediately painted The Shriek or The Shout or The Squawk, just with exponentially more pieces of art referenced. In fact, almost this exact scenario is boasted about on the OpenAI website, where Johannes Vermeer’s Girl with a Pearl Earring is “reimagined” numerous times, all in the same art style as the original.
This is all to say nothing of the fact that the Internet is, well, the Internet, and that DALL-E 2 is sourcing its data from the variety of reference material to be found there. As a result, damaging social biases, inappropriate materials, and other potentially problematic content could easily find its way into the images of DALL-E 2—and by extension, Microsoft’s integration of it.
However, it is important to note that Microsoft has been communicative about how these issues are being proactively addressed, citing their partnership with OpenAI to ensure that safeguards are put in place to prevent the aforementioned concerns and demonstrably taking their time with the rollout of these tools.
If not, what are a few things that your business could use to improve its technology? Give us a call to find out how we can make it easier for you to implement and maintain these tools and solutions. Reach out at (503) 278-5011.