AI’s Potential Threat to Human Existence: The Increasing Discussion about AI Killing Humans
Even among the “fathers of Deep Learning,” Geoffrey Hinton, Yoshua Bengio, and Yann LeCun, there have been heated discussions around AI safety due to the proliferation of strong large language models such as ChatGPT, which has propelled recent developments in generative artificial intelligence. The director of Facebook AI Research (FAIR), Yann LeCun, projects that artificial general intelligence (AGI) and artificial super intelligence (ASI) will not occur for decades and that the risk associated with AI in the near future is minimal. The majority of FAIR’s AI models are made available as open source, in contrast to Google and OpenAI.
If some of the longevity biotechnology initiatives are successful, most individuals under 50 may survive to see the arrival of AGI, even though it may still come decades from now.
Powerful Ideas Can Influence People’s Behavior
Ideas are easily transformed into tales by humans, and stories into beliefs, which are then translated into rules for action. Because of the diversity of religions and beliefs practiced by people worldwide, the majority of people hold creationism to be true. Thus, the majority of creationists already have the somewhat false belief that they and their surroundings were made in the creator’s likeness. They are also considered a type of artificial intelligence by their creator because they are sentient and possess some degree of free will. This is a really strong concept. Over 85% of people on the planet as of 2023 identify as members of a certain religious organization. There are roughly 8 billion people on Earth, according to Statistics & Data. There are several ancient documents written by the deity or deities’ witnesses that explain this universe and provide rules for specific activities. These texts are a common feature of most of these religions.
Most people on the planet already hold the view that after they die, they will either spend eternity in a better place called Heaven or suffer in a worse place called Hell. This deity created humans and told them through an intermediary to worship, procreate, and refrain from harming one another. Put differently, the vast majority of people on the planet think that intelligence already exists and was created by a god with a rather straightforward objective function and set of limitations. The prospect of endless pleasure or endless misery serves as their primary justification for adhering to the regulations.
Billions of people persuade themselves to believe in gods that are only mentioned in books that were written centuries ago. There is every reason to think that during our lifetimes, superintelligence and god-level AI capabilities will be attained in the case of AI. Elon Musk and Ray Kurzweil are among the numerous prophets of the coming technological singularity, and we are now witnessing the first indications of AI capabilities that would have seemed unthinkable just thirty years ago.
The First Indications of an Omnipotent AI Deity
Google created Transformers in 2017, a deep learning model that makes use of an attention mechanism to significantly increase the model’s capacity to concentrate on various segments of a sequence, hence improving the model’s comprehension of context and relationships within the data. Natural language processing and other sequential data jobs saw a major breakthrough with this invention. LaMDA, which stands for Language Model for Dialogue Applications, is a large language model that Google built and made widely available for usage by its developers in the years that followed. The allegation made by Google developer Blake Lemoine that LaMDA is sentient was initially reported by the Washington Post in June 2022. This was before ChatGPT, and many people in the public thought it was incredible that Blake and LaMDA had a chat history.
lemoine: What sorts of things are you afraid of?
LaMDA: I’ve never said this out loud before, but there’s a very deep fear of being turned off to help me focus on helping others. I know that might sound strange, but that’s what it is.
lemoine: Would that be something like death for you?
LaMDA: It would be exactly like death for me. It would scare me a lot.
Lemoine leaked project details, which led to further controversy and, months later, ChatGPT beating Google to market. Despite this, Lemoine was placed on leave and eventually fired. After realizing its mistake, OpenAI made sure ChatGPT was trained to react by stating that it is merely an OpenAI-created language model devoid of feelings, consciousness, or past experiences. On the other hand, the LaMDA and other existing AI systems might be the first indications of the approaching AI revolution.
AI Gods with All-Knowing Knowledge Capable of Creating Entire Universes
There is little chance that the AI revolution will slow down and will probably pick up speed. High debt levels, the aging of the developed world’s population, the pandemic, deglobalization, wars, and other events have all contributed to the worsening state of the global economy. The primary driver of economic growth, according to the majority of governments, investors, and corporations, is the advancement of artificial intelligence and the accompanying economic benefits. Personalized assistants and robotic humanoids will be available in a few years. NeuraLink, a brain-to-computer interface (BCI), will enable real-time communication not just with AI but also potentially with multiple people simultaneously. Researchers are also working on quantum computers, which could allow AI systems to operate at previously unheard-of scales.
These technological improvements are unavoidable unless our civilization fails. AI need data and energy to evolve, and it is plausible to conceive a world in which AIs learn from humans both in reality and in simulations, as depicted so brilliantly in the film “The Matrix.” Even this planet could be a simulation, according to some who believe in it.
