Original link: https://www.indigox.me/if-we-dont-control-ai-ai-will-control-us/
This article is excerpted from the latest article ” You Can Have the Blue Pill or the Red Pill, and We’re Out of Blue Pills ” in The New York Times｜Opinion column. I made some conclusions and personal comments on the article. I want to get For an authentic reading experience, please read the original text!
Original author: Yuval Harari (historian, author of a brief history of mankind, founder of Sapienship International Consulting Company), Tristan Harris and Aza Raskin are the founders of Center for Humane Technology , and they are also the producers of the well-known documentary ” The social dilemma “.
Harari and Harris show a forward-thinking concern in the same way that pharmaceutical companies cannot sell new drugs to people without going through strict safety checks; AI systems with GPT-4 and above capabilities should not be faster than the masses To the extent acceptable to the culture, put its function into full use. OpenAI and Google should be controlling the release rate of AI so that humans do not have difficulty adapting. .
1/8 It is difficult for our human minds to comprehend the new capabilities of GPT-4 and similar tools, and even more difficult to understand the exponentially more powerful capabilities of these tools. Most of the key skills come down to one thing: being able to manipulate and generate language! Language is the operating system of human culture . Our myths and laws, God and money, art and science, friendship and countries, and computer code all derive from language. The mastery of language by artificial intelligence means that it can now hack into and manipulate the operating system that represents human civilization. By mastering language, artificial intelligence is seizing the fire of civilization 
2/8 What would it mean if humans lived in a world where most of the stories, music, images, laws, policies, and tools were shaped by non-human intelligence? This intelligence knows how to exploit the weaknesses and biases of the human mind and how to develop a close relationship with humans. At the most advanced game of Go, no human can beat a computer. What happens when this happens in art, politics or religion? AI could quickly swallow the entire human culture we’ve created over millennia , digest it and start to spring up a host of new cultural artefacts, political speeches, ideological manifestos, or the holy books of cults. I am quite sure of this!
3/8 Human beings generally do not have direct access to reality, we are surrounded by culture and experience reality through a cultural prism . Humans have been living in other people’s dreams for thousands of years. We worship gods, pursue ideals of beauty, and dedicate our lives to causes imagined by some prophet or statesman. Soon, we, too, will find ourselves living under the illusion of non-human intelligence. . .
4/8 Just by mastering language, AI can have all the necessary conditions to trap us in a fantasy world similar to “Matrix”, without needing to shoot people like “Terminator” or implant any chips into in our brains. If a shot is necessary, the AI can tell the story to get the human to pull the trigger.
5/8 social media was the first contact between AI and humans , and humans failed. This contact let us know what ” The social dilemma ” is. In social media, raw AI is being used to sift through user-generated content, not create it. Primitive though it is, the AI behind social media is powerful enough to create a curtain that increases social polarization, damages our mental health and unravels democracy. I wrote this article ” From Social Dilemma to Social Dilemma / Rules of Survival Under the Digital Wall ” in early 2021.
6/8 Large Language Models (LLMs) are our second touch with AI . But is there any reason to believe that humans are capable of aligning these new forms of AI with our interests? If we continue to do things as usual, even if it inadvertently undermines the foundations of our society, new AI capabilities will once again be exploited for profit and power.
7/8 should think hard about AI before our politics, economy, and daily life become dependent on it. Democracy is a conversation, and the conversation depends on language, and when the language itself is hacked, the conversation collapses and democracy becomes untenable. If we wait for the chaos to happen, it will be too late to correct it. When godlike power is matched with the control of corresponding responsibility, we can realize the benefits that AI can promise. This issue was brought up in a conversation between Elon and Sam in 2016.
Elon Musk discussing the dangers of centralized AI with Sam Altman (CEO of OpenAI), in 2016.
Fate loves irony. pic.twitter.com/SloROu48mb
— Tesla Hype (@TeslaHype) March 25, 2023
8/8 AI it’s like we have summoned an alien intelligence we know very little about except that it is extremely powerful but could also destroy the very foundations of our civilization our social structure and governance were born in the 19th century The industrial age is clearly not ready for this coming intelligent age , this architecture must be upgraded, and we must learn to master AI before AI masters us 
This “If We Don’t Control AI, AI Will Control Us” from historian @harari_yuval and HumanTech co-founder @tristanharris shows a forward-looking concern that pharmaceutical companies cannot Same as selling new drugs to people before checking; AI capable of GPT-4 and above https://t.co/icZcJrMTVP …
— indigo (@indigo11) March 26, 2023
The above content comes from my Twitter. For a better reading experience, I will share the Twitter threads of excerpts and reflections into blog posts and mailing lists!
This article is transferred from: https://www.indigox.me/if-we-dont-control-ai-ai-will-control-us/
This site is only for collection, and the copyright belongs to the original author.