How AI artwork software could create legal minefield, supercharge misinformation online

If the technological wonders of the AI writing tool ChatGPT weren't enough for you, similar technology can be used to create artwork, photography, or soon, even a perfect copy of a human voice.  

But experts warn it could create a legal minefield and supercharge misinformation online.

"The technology behind ChatGPT could make a bigger impact on society than the smartphone has done," said Paul Spain, Gorilla Tech CEO. "Let's see how that plays out but it's totally possible."

For example, you can give these platforms a simple prompt like "An oil painting of a boy in a boat in the style of Monet" - and it will deliver you results in seconds.

It's transformative but also a potential legal minefield.

"It opens a can of worms," said lawyer Arran Hunt.

The problem is that AI content generators are sometimes trained using existing works, often without the artist's consent.

"I think we need to start looking at updating our law, perhaps the Copyright Act, as to where we can draw inspiration from," Hunt said. "Currently the copyright exists in the piece itself but not in the style of the piece."

But copyright issues aren't all that's at stake. AI-generated content could soon be indistinguishable from official sources and that ramps up the threat of misinformation. Experts said regulators must move faster than with the previous tech.

"We have a moment now where we are aware of it," said Dr Sarah Bickerton, Research and Policy Lead, Tohatoha Aoteroa Commons. "We need to start putting policy in now in the way that we didn't for smartphones, for algorithmic search functions, for social media."

But this story doesn't stop with words and pictures. Microsoft is currently developing an AI voice creation tool, VALL-E, which from just a three-second prompt of a human voice can clone that voice to say anything at all.

But it's the combination of these tools, which could be truly revolutionary and frightening. Soon, by using a text, image and voice-generating AI together entire news stories - like the one in the video player above - could be created with virtually no input from a human.