Ghibli style AI

The last few days have been an Internet frenzy on the ‘Gbhili style art’. I have noticed that many of us are on either side of the spectrum here. Either, we are too excited and want to learn about it and join the bandwagon right away or we are on the fence and waiting for ‘Ghibli dust’ to die down so that they can take their own decisions after that(by when I think another ‘Internet sensation’ would have come into existence and ‘Ghibili art will become a sweet memory! :))

I always belong to the former category and have to know what is trending in the tech world by tinkering with it! 🙂 It has been a fun ride and the pictures generated in the ‘Ghibli style’ have been looking smooth and nice.

There are two sets of worries from what little I could see:

I. Privacy concerns:

The pictures were encroaching on people’s privacy and uploading personal pictures to any platform that generated Gbhili style images will store them and misuse them and use it to train their AI models. This is a perfectly valid concern. I am a security and privacy professionals and have learnt a lot about privacy concepts in the last few years. I would never encourage anyone to upload their personal pictures, personal details to any platform under any circumstances. It is a definite possibility that pictures can be stored to train AI models. However, there are times when we do get carried away and want to have some Internet fun! 🙂 This might be one such case of people having a little bit of fun…..

In all seriousness, privacy laws are getting stronger each year and there will be suitable settings in all AI models to turn off ‘AI model training’. It is up to us to read the privacy settings and make sure that the appropriate boxes are checked off(to prevent OpenAI, Grok or any other platform form using your information to train their AI models) This can be done in ChatGPT as follows:

Click on ‘Settings’ -> ‘Data Controls’ -> ‘Improve the model for everyone’ as ‘Off’

So, we have addressed one concern. A word of caution though – in spite of ensuring safe security and privacy settings, it is good to use all tools with caution and leave the children out of this.

II. Copyright issues

The second concern was stealing Hayao Miyazaki’s (the creator of the Ghibli style artwork) and AI recreating their art works. Unfortunately, in the AI age, we all seem to be just puppets in the hands of AI puppet master. According to the Hindustan times, “The Ghibli style art work is not protected and OpenAI is not legally violating any laws”(Source: https://www.hindustantimes.com/world-news/using-chatgpts-ghibli-style-photos-violates-copyright-laws-experts-explain-101743257416332.html) So, as of today, OpenAI or Grok or any other AI model is taking a conservative approach to protect the artists work and allowing users to generate only a few number of Ghibili style images.

I think in the AI age unless, somebody explicitly states that their work is copyrighted and cannot be used by anybody or AI, it can unfortunately be used.

Do you have any creative work that can be (mis)used by AI? Better copyright it and seek safety from AI models as the future looks bright for AI model training.

This post is for alphabet ‘A’ for BlogchatterA2Z marathon by Blogchatter!