Artificial Intelligence and images

Artificial intelligence is constantly evolving and images are, of course, involved.

The launch of DALLE-2 (an Artificial Intelligence (AI) programme that creates images from text captions) raises questions on how such technology could impact illustrators, with both short and long-term effects. Downsides may include Art Directors who want to reach for a fast and potentially lower-priced illustration accessing AI platforms, particularly if they are not looking for something of exceptional quality. This raises similar issues to the use of lower priced stock imagery for a project rather than commissioning an original work from an illustrator.

However, it’s important to acknowledge an illustrator’s input to a commission starts long before delivering the final image. Illustrators are experienced in supporting clients with ideas generation and exploring the boundaries of a brief to create the best – or surprising – results. Should AI image generation become a popular tool for commissioners, illustrators and art directors risk losing this invaluable collaboration and development process which can hugely benefit the client.   

Illustration often relies on human experience to communicate an emotion or opinion and connect with its viewer. Something that’s yet to be proven is whether DALL-E 2 or other platforms would be able to grasp concepts and reactions to current affairs, political events and social issues, or set the right tone needed for a narrative, such as a children’s book or graphic novel. Many topics require sensitivity when creating visuals to illuminate them, and this too may be a challenge for AI image construction.

Presently there have been some discussions around copyright in the images used to ‘teach’ AI, and how that should be compensated to the image rights holder. Overall, the issue of creating really good work would depend on ‘feeding’ it, or ‘teaching’ it to use high-quality work from real illustrators and photographers. This can be done by ‘data mining’ – machine analysis of online content. This process brings up the question of appropriation of styles, copyright permissions, and the compensation for use of images.

Florian Koempel, director of IusArtium, a consultancy for the creative industry, emphasises the importance of financial remuneration for creators in his piece Artifical Intelligence and copyright – superficial observations

‘In all the excitement about the opportunities created by Artificial Intelligence we must ensure that the human creators and artists are properly remunerated for their talent and work, collectively and/ or individually.  Creators and artists already employ new technologies, but their endeavours – training the machine or creating with their help – need to be remunerated.’

Currently the law allows ‘mining’ for the purpose of machine analysis of online content for non-commercial research by users who have legal access to the content, such as subscribers. However, in June 2022, the UK government announced that they will introduce a new copyright exception for Text and Data Mining that would allow anyone with lawful access (a phrase that has not been defined) to mine content for research, but also any commercial usage. In the government consultation that had preceded this decision, organisations for creatives had argued that mining should remain for non-commercial research purposes for those with legal access. AOI was part of the British Copyright Council submission for the consultation.


7th July 2022
Share

Issue 42 and Back issues available!
Contacts (across 3 sectors) for 350+ UK Commissioners... BUY NOW!
Get seen by commissioners with a 20 image Folio.

JOIN OUR MAILING LIST

Join our mailing list to get the latest news.