Between innovation and ethical dilemma

Sep 23, 2024  •  4   • 697 

Until now, I have avoided writing about the topic of ‘AI’. This was a conscious choice, since I feel that everyone is already writing about this and you almost get ‘AI tired’ by now. Nevertheless, in a short period of time I came across a number of things that inspired me to write down my thoughts on the subject. Think of it as a diary text, nice to read back years later (or in this AI era, after a few months) and then look: ‘wow, that was still old-fashioned, they should have known back then…’.

Trends

The introduction of all AI tools is moving incredibly fast and can hardly be kept up with. So I let go of that ambition, while Google came out with the NotebookLM I was triggered to try it out. With NotebookLM you can very easily upload sources, post text and within minutes a podcast is created. The first time I heard this I knew it was an AI produced podcast, despite that, I was very impressed. The second time I heard from the voices and style that it was a generated podcast. Currently only available in English, but I guess it’s just a matter of time that it will become available in more languages and with more voices. So it will become increasingly difficult to detect whether it is an AI podcast or a “real” podcast.

I uploaded the text of this blog and had it made into a podcast. This takes me 10 seconds to upload and after a few minutes of waiting I had this podcast.

“The process is different.”

This brings us into interesting territory, what is ‘real’. My initial response was that it puts people on the wrong track when you have them listen to an AI-generated podcast, I think it is wrong if that’s not communicated to the listener. The backlash I received was, “People are listening to something real. It’s a fallacy that using a tool suddenly doesn’t make something real. The process is different.” This included a comparison to modern music, where there are AI-based components in it, both on the sounds, mixes and lyrics. “Just like with synthesizers, initially people didn’t think that was real music, but it was just a new tool that can be used to make music.”

Is a podcast different from visual material?

This raises many questions:

• Do we need to call it artistic freedom?
• What makes something real?
• Is it important for something to be real?
• What about source references?
• Is it indeed just a new tool?
• What if the tool was ‘hallucinating’ during creation?

I personally find authenticity and the possibility of fact-checking particularly relevant. With a radio play on the radio or a cartoon, we all know it’s not ‘real’ either. A podcast people listen to to learn and increase knowledge, is it then justified to put resources into an AI tool and have AI create the podcast and then put that live as a ‘self produced podcast’ or is it sufficient to mention that it was created by AI?

Ethical dilemmas

The rapid development of AI tools, including for podcast creation, requires thorough ethical reflection. It is important to think about transparency, impact on listeners, artistic boundaries, accuracy of information and the potential dangers of plagiarism. Let’s start by raising awareness and not taking advantage of ignorance among many people less involved in recent IT developments.

Dear diary

Dear diary, I’m not quite sure what to make of all the AI and switch between “oh look how nice,” to “quite an exciting development too". I am very curious how we will view all these developments in the future. It can be fantastically helpful in programming or as a source of inspiration and yet I do see some red flags. I can say that the rapid developments in AI are creating a special and sometimes frightening time with both benefits and risks.

Recommendations

• 📚Books:
  • “Co-intelligence: Living and Working with AI” (2024) (in English)
  • “Echt nep” (2021) (in Dutch)
• 📻 (real 😊) podcast (in Dutch): “Poki – Podcast over kunstmatige intelligentie”

Disclaimer: this blog is written by myself, without the use of AI. Opinions expressed are my own.