how can ai overcome copyright infringement issues?
by rick richardson
technology this week
adobe has unveiled firefly, a new suite of generative ai tools it hopes will encourage more design professionals to embrace rather than fear this cutting-edge technology.
why it matters. as generative ai has swept the globe, artists have had difficulty adapting to it. while some view it as a useful tool, others see it as a threat to jobs and a kind of thievery.
(wait, where have we heard that before?)
more tech this week: chatgpt: what it is. why you need to know. | russia-linked ransomware back with a vengeance | amazon aws: the mainframe killer? | amazon launching its first internet satellites | russian solarwinds hackers at it again | nasa finds a $10 quintillion asteroid | firms must balance benefits, risks of emerging technology | microsoft and google go to war
exclusively for pro members. log in here or 2022世界杯足球排名 today.
adobe wants to address two obstacles that are preventing business usage of the technology with firefly: worries about copyright issues and the absence of professional-grade tools.
- on top of adobe’s extensive font library, the initial firefly model provides both a text-to-image engine and the ability to create visual text effects.
- besides pictures from adobe stock, the model was trained on pictures that were in the public domain or distributed under an open license.
- although adobe hopes to incorporate firefly into its family of creative tools, starting with adobe express, photoshop and illustrator, as well as the adobe experience cloud, it is initially only available in beta form on the web.
there is intense disagreement within the art community about how to react to introducing generative ai tools like stable diffusion and openai’s dall-e 2.
a few of artists who embrace technology the most are already using generative ai in their creations, including some whose work has been on show in a san francisco gallery.
stable diffusion has been sued by getty images, which claims that its engine was inappropriately trained on copyrighted images.
training the ai system on stock imagery offers adobe a couple of advantages for avoiding intellectual property infringement.
- compared to search engines that are trained on content from throughout the web, adobe has more rights and control over images from its adobe stock database.
- brand names and logos, as well as other forms of copyrighted content, are frequently absent from stock pictures.
according to adobe, anyone that submits their work to stock will be able to take advantage of generative ai’s revenue opportunities. but the corporation has provided no information about how this will operate. david wadhwani, president of adobe’s digital media division, summarized by saying, “we want it to be safe for creative usage.”