The new AI Text to Vector Graphic (beta) Feature Sucks

This reminds me of the early Jive days when it was impossible to post anything mentioning the Live Paint Bucket (Illustrator tool).

Hi @POWDERMONKEY71,

I'm sorry to hear that you're frustrated with the Text to Vector Graphic (beta) feature in Adobe Illustrator. While AI and machine learning technologies have made significant progress in generating content, they are not without their limitations.

However, the inconsistency you've described could be due to a range of factors, including the specific keywords used in the prompts. While it might seem arbitrary, machine learning algorithms use complex models and large datasets to make these determinations.

Your feedback is valuable, and it helps us understand how to improve the feature. If you're open to it, we would appreciate it if you could share the file you were working on and provide details about the specific prompt or text input you used. This information will allow us to take a closer look at the issue you encountered and potentially make enhancements to the feature.

Additionally, within the Text to Vector Graphic Beta, there is a feedback option. Feel free to use it to share your thoughts directly with our team. Your insights can contribute to the ongoing development and refinement of the feature. We value your input and are committed to making improvements based on user feedback.

Thank you for helping us enhance your experience with Adobe Illustrator!

Best,

Anshul Saini

no problem with Axe in spanish either

It's inconsistent. "Axe" is being blocked.

"Hatchet" or "Hatchet and big axe" or "Axe – a tool for a craftsman" don't get blocked.

"Arbalest" gets blocked. "Crossbow" is accepted.

Looks like sometimes you have to game the system.

Just out of curiosity I had it generate a chainsaw and an axe for me. And it would happily do that. One of the images it created, contained an axe that was separate from the chainsaw. You could also ask for an axe and woodworking tools. It did that as well for me.

Have you been keeping up with current events regarding this "AI" stuff? This is leaps and bounds beyond the warnings about photo manipulation when Photoshop was in its infancy 30-plus years ago. Back then it took both time and considerable skill (not to mention serious money for halfway decent computing hardware) for someone to generate a convincing looking fake photo to use for illegal purposes.

AI-driven features in graphics software could make nefarious tasks easier for people who don't really have any real skills using these applications. Combine that with America's hysterical political climate. Those conditions should make it easy to understand why Adobe had to incorporate at least some guard rails. They don't need the US Congress telling them what restrictions to incoporate.

Some of the restrictions are a bit much. Just for a goof I tried using Photoshop's Generative Fill feature to cover a blank canvas with a vomit texture. I thought it would be funny. Photoshop refused to do it. I get it.