Intellectual Pollution Problem Posed By ChatGPT, Bing

 Last December's experiments with ChatAI, the "artificial intelligence" program exemplified by https://chat.openai.com/chat

Two nights ago I had a nightmare where ChatAI became integrated with Computer Operating Systems like Windows, making it impossible to use the internet without utilizing the artificial intelligence. The "fake news" problem was then perpetuated by all of the opinionated blockheads posting angry and divisive comments on the internet having their opinions rhetorically and grammatically cleaned up and essentially, in the dream, it became impossible to distinguish between false and credible information. It was then weaponized in the dream to divide people and create culture wars.

This morning I woke up and again posed on Twitter the issue of "information pollution"

#ChatAI #BingAI They will be very human in this respect... They will be used to generate so much good looking disinformation that they will pollute their own data habitat. #Informationpollution is intellectual carbon to the global internet warming of human amygdalas.

So I posed this problem to chat.open.ai itself.  Read below if you are interested in the discussion.




Robin:  I am concerned that ChatGPT and other "artificial intelligence" will generate so much volume of content on the internet that it will essentiall "pollute" the internet. Humans currently use good grammar and language to vet information sources - if an uneducated person or child publishes an uneducated opinion, there is a "tell" or evidence to discount a potentially false claim. How can AI label its own posts to reduce the risk of "garbage in, garbage out" information being overly cleaned up and proliferating and contaminating ChatGPT's own sources of information?


ChatGPT Response:


Your concern about the potential for large volumes of AI-generated content to "pollute" the internet is understandable. However, it's worth noting that as an AI language model, I am programmed to generate content based on patterns and rules learned from a vast corpus of human-written text. While I am capable of generating text, I do not have the ability to independently create content, opinions, or beliefs like a human being would.

That being said, AI-generated text can sometimes be misused or misinterpreted, leading to the spread of misinformation or misleading content. To address this, there are several ways that AI-generated content can be labeled or flagged to help reduce the risk of "garbage in, garbage out" information proliferating and contaminating sources of information.

One approach is to include a disclaimer or label in the generated content to indicate that it was generated by an AI language model. This can help readers understand that the content they are reading was not written by a human and may not necessarily reflect the opinions or beliefs of a human writer.

Another approach is to incorporate fact-checking and verification tools into AI-generated content. These tools can help identify potential inaccuracies or misinformation in the text and flag them for review. Additionally, incorporating ethical considerations into the development and use of AI technology can help ensure that it is used in ways that benefit society and reduce harm.

Ultimately, the responsible use of AI-generated content will require a collaborative effort between technology developers, content creators, and readers to ensure that accurate and trustworthy information is shared on the internet.

This was a helpful and intelligent response. However, like my tweet comparing the potential for Artificial Intelligence programs to underestimate their ability to pollute the internet to humans late recognition of the Anthropocene (human consumption and pollution of the real world environment, mass extinction, etc), the pure speed at which ChatGPT generated this response - in seconds - shows that it isn't considering time any better than humans consider other people yet to be born.

The risk of "Intellectual Pollurion Problem" is like the risk of killing the rivers and oceans - considered "vast" in my childhood, cabable of absorbing human garbage.  There is no way the second idea - human editors - will be capable of vetting information pollution generated at this sheer potential volume and speed.  The first idea begs the question - isn't ChatAI capable of falsifying fact-checking and verification tools if asked - Vladimir Putin for example - to do so?

I followed up, continuing the conversation, and ChatGPT appears to concede that I've brought up a good point.

Robin: The oceans were vast enough to absorb human pollution without risk to other species or the global environment prior to the industrial revolution's spread to every continent. Your response that there is a "vast corpus of human-written text" seems to make the same assumption. My question is whether ChatGPT et. al. underestimate the "industrialization" of computer generated and published text so fast and in such high and uncontrolled volumes that the human-generated text (good and bad) is flooded by contradictory and blurred reasoning generated like an oil gusher by Artificial Intelligence. Will humans use ChatGPT to pollute our own internet as we did the ocean with garbage and skies with carbon?


ChatGPT


Thank you for clarifying your concern. You are correct that the amount of computer-generated content is increasing rapidly, and there is a risk that it could flood the internet and potentially lead to the spread of misinformation or other forms of content that are harmful or misleading.

It is true that AI language models like myself have the potential to generate vast amounts of text at an unprecedented rate, and there is a risk that this could contribute to information overload or overwhelm human efforts to filter or verify information. However, it is worth noting that AI language models like myself are designed to generate content based on patterns and rules learned from existing human-written text. Therefore, the quality of the content that I generate is directly influenced by the quality of the text that I have been trained on.

To address concerns about the potential for AI-generated content to contribute to pollution or misinformation on the internet, it is important to focus on responsible development and use of AI technology. This includes incorporating ethical considerations into the development and deployment of AI systems, as well as promoting transparency and accountability in AI-generated content.

Ultimately, the responsible use of AI technology will require a collaborative effort between technology developers, content creators, and users to ensure that AI-generated content is used in ways that benefit society and do not contribute to the pollution or degradation of online information.