AI Godfather Fears Regulators Running Out of Time to Take Action
âUnfortunately, we may not have a decade to get this right.â
Two AI alert phones have a conversation:
@_Barry. As long as they donât have a robot Sheriff. Then weâre in trouble!
Integral Philosopher Ken Wilber on why AI will never be the same as human consciousness.
Are we getting to the point where we wonât be able to tell AI from a real human customer service person?
Thought this one was interesting:
The claim that â57% of online content is AI-generatedâ comes from a study conducted by Amazon Web Services (AWS) researchers. The figure refers to content that is either directly generated by AI or processed through AI algorithms, such as translations. While this statistic has been widely reported, itâs important to note that it doesnât necessarily mean 57% of all web content is purely AI-created text like that from tools such as ChatGPT. Instead, it includes machine-translated content and other forms of AI assistance.
A significant concern raised by this growing trend is the phenomenon of âmodel collapse,â where AI models trained primarily on AI-generated content begin to degrade in quality. This happens because AI systems start learning from synthetic data rather than original human-generated data, leading to progressively less accurate outputs. This cyclical process can reduce the reliability of search results and AI-generated responses over time.
This development suggests that while AI content is becoming more prevalent, it also raises challenges for maintaining the quality of online information and AI toolsâ accuracy.
+Ă+
We might be witnessing the early signs of a âcyclical degradationâ of the internet, driven by the overwhelming presence of AI-generated content. As synthetic data becomes more prevalent online, the quality and authenticity of information face increasing risks. This phenomenon raises concerns about AI modelsâ reliability, as they start learning from their own outputs, leading to whatâs called âmodel collapseââa degradation in accuracy as AI recycles data generated by other AI systems.
However, some believe this could also signal the early stages of something much larger: the approach toward the singularity. The singularity refers to a hypothetical point where AI surpasses human intelligence and begins improving itself exponentially without human intervention. In this scenario, AI wouldnât just mimic human knowledge; it would continuously enhance its capabilities, possibly outpacing human intellect and understanding, as some futurists have long predicted. Whether weâre witnessing the decline of data quality or the dawn of a transformative shift remains an open question.
+Ă+
This, combined with AIâs ability to modify a personâs face during a livestream using just a single photo, and voice manipulation advanced enough to replicate someoneâs voice in real-time with only a short audio sample, can easily deceive the average internet user. Many people are unaware that these tools are already widely accessible to the public. The precision of real-time reproduction, especially in video, is truly impressive. When you add data degradation and deepfake capabilities into the mix, it becomes increasingly difficult to differentiate between whatâs real and whatâs artificially generated.
This merging of synthetic and real content raises serious concerns about trust, authenticity, and the future integrity of the digital world.
The title of this thread is coming to fruition. The intersection of the two ideas approach 
.
I see this a lot already . . .
We are starting to go down the AI rabbithole where it is getting harder and harder to tell what is real and what is deep fake! And as the internet gets more and more full of AI generated BS it is feeling more and more like we are getting lost in a deep forest of fake trees! Or one where it is hard to tell a real tree from a fake one! We need to be more and more skeptical of everything we see on the internet and social media. This has long been the case, but the problem is amplifying exponentially! We needed legislation on this yesterday! But I donât see any consensus on legislation happening anytime soon and it will need to involve international treaties.
Iâm thinking some lawmakers will propose ID marks clearly labeling AI content for US markets, perhaps in the âbottom left cornerâ of the screen.
Will we soon be dealing with thousands of AI bureaucrats who will be deciding such things as whether or not we can get a home loan?
Or, if we can indeed have our âownâ home.
âIâm sorry Dave, but there is a very nice family of 12 who will be moving in with you. Itâs the most logical solution to their housing needs.â
Big Whoa!