Skip to main content
Adobe Employee
March 22, 2023
Open for Voting

If two people add the same prompt, how to ensure that they are getting different results?

  • March 22, 2023
  • 4 replies
  • 489 views

With so many people using the platform, it might be possible that people end up using the same prompts and would like to use the image for their work. How are we ensuring that these pieces of art are different and would not create copyright issues? 

4 replies

Participant
May 31, 2023

I think AI tools shouldn't be used for profit/ commercial use; it is a symbiotic entity that reflects the mind of all humanity.

 

Just as we insert the same prompts, we also insert the data sets leading to their resulting work.

 

Maybe, if it needs to be asserted, a new form of licensing  should be created[for AI assisted works due to circumstantial need (time, concepts/drafting disability)] 

A stipulation could be part of the profit is aggregated to a charity or fund for AI servers, research, or another equivalent.

 

Then, I guess, that creates a dilemma within tax deductions [North America, USA]

 

The ethics of everything will inevitably be innovated. These topics revolve around capitalism, proprietorship, and much more.

Wizzerd L.A.
Adobe Employee
March 22, 2023

Interesting, would love to learn more about how we are providing safety against piracy and copyright issues in the upcoming days. 

Another-Phil
Inspiring
March 22, 2023

Hi @strandsofcreativity  - Once Firefly is generally available, it will be designed to be safe for commercial use. Images created during the beta will have a watermark indicating they are not for commercial use.

Fabio Staiano
Participating Frequently
March 22, 2023

It's precisely the same thing I asked on Discord a few hours ago. 

Hard to find a solution. I made a few tries to understand the proximity of results, and even if it's not exactly the same result, it're pretty similar. Way more than other AI tools - I guess because of a smaller dataset for ethical reasons.

I think the only way out is to expand the existing dataset, to make repetition statistically impossible.