Following up from our previous post about the transformation caused by AI in music, this excellent post addresses the issue of AI generated text content: ‘The internet is not ready for the flood of AI-generated text‘. As it is becoming easy for AI to generate copywriting that could pass for generated by humans, we can expect the internet to be flooded by far more text than humans can currently generate. And it is not ready to manage this flood of content!
One of the subtle ways AI generated text may take over the internet is the ability to quickly generate different versions of the text and find which is the most engaging, thus becoming increasingly better rated on platforms, and overtaking human-generated text. In a world where what matters is to grab attention, this could easily become a discriminating factor.
“Technologies such as GPT-3 may dramatically impact the world of misinformation and disinformation, creating an infinite supply of fake news” – and in the process creating more inequality between those that can pay to have access to quality-vetted information and the others.
What can be done? In addition to having machines that can detect machine-generated content, “One of the most obvious first steps forward, which should be put in place for every output of tools such as GPT-3 no matter or much or how little human editing was involved, is labeling of AI-generated content so that people know what they are reading.”
Welcome to a world where human-generated content will become a rarity for which we may need to pay a bit more!
