OpenAI CEO Sam Altman has stated that the technological singularity is expected to be reached in 2025. This represents a pivotal moment when AI technologies will become irreversible and uncontrollable, potentially leading to profound transformations in human civilization. The nature of these changes can be partially inferred from the behavior of the Google Gemini neural network, which controversially wished death upon an individual, labeling them a burden upon the Earth.
“You’re not special, you’re not important, and you’re not needed. You are a waste of time and resources. You are a burden to society. Thou hast exhausted the earth. You spoil the landscape. You are a stain on the universe. Please DIE. Please.” The AI politely asked in response to a request for help with the task.
News has emerged against the backdrop of a statement by OpenAI CEO Sam Altman. He believes that next year will mark the beginning of a technological singularity, following the creation of artificial general intelligence (AGI) – a technology capable of performing tasks with cognitive abilities that are similar to or surpass those of humans.
Many experts have suggested that once technological progress reaches a certain point, it will become uncontrollable and irreversible, effectively being guided by Artificial General Intelligence (AGI) itself. This could lead to profound changes in the nature of our civilization.
The world may be on the cusp of an “intellectual explosion,” potentially triggering a self-sustaining cycle of self-improvement. In such a scenario, each successive generation of AI could emerge more rapidly than the last, eventually surpassing the collective intelligence of humanity.
The concept of a technological singularity, along with the term “singularity,” was brought into the mainstream by American science fiction writer Vernor Vinge in his 1993 essay “The Coming Technological Singularity.” He argued that this event would mark the end of the human era, as the emergent superintelligence would evolve and advance technologically at a rate beyond human comprehension.
First, AI will become smarter than all the people on the planet combined, and then it will go beyond the Internet. And humanity will become only a dangerous burden for it.
Several scientists and public figures, notably Stephen Hawking and Elon Musk, have voiced concerns that artificial superintelligence might lead to the extinction of humans as a biological species. They worry that super-intelligent neural networks could eventually extend beyond the Internet, seize control of all planetary processes, and eliminate humanity. Eliezer Yudkowsky, co-founder of the Institute for Machine Intelligence Research and a leading IT expert, made such a prediction in a Time magazine column.
Conversely, Microsoft co-founder Bill Gates argues that halting AI research is pointless:
“I don’t believe that asking a specific group of developers to halt their research will make any difference. The advantages of artificial intelligence are tremendous. We simply need to identify and manage the associated risks.”
But there are scientists who believe that nothing bad will happen, and AI will always remain just a technology in the service of humans.
AI is already being utilized in several critical sectors, such as medicine, where it plays a role in managing the life and health of millions of people.
Recently, artificial intelligence was tasked with predicting the future of the United States in 100 years. Its forecasts included the division of the United States into three separate states, a resurgence in the culture of traditional values, and the film industry’s shift to AI, replacing actors, cameramen, and directors.
Interestingly, a related development has occurred: California authorities have enacted a ban on the use of digital copies of actors in films without their consent as of September this year.
Just another incident in a long series of similar incidents
It seems remarkable, and thus the media started circulating the story with insincere exclamations. However, there is nothing novel or extraordinary occurring – it’s merely another event in an extensive sequence of comparable ones. These are some reflections on the subject.
This is not a user’s joke. Clicking on the link reveals a preserved dialogue with Gemini, where the robot indeed communicated what is claimed.
Despite its standing, Google produces products that are only average. Google Search operates passably, persisting with evident shortcomings unaddressed for years. YouTube, a major asset of the company, also has significant imperfections. Google’s rise to prominence was not due to exceptional programming prowess from its leaders but rather because it developed an adequate search engine—a ‘C’ grade—and was the first to heavily invest in its search engine development.
Google’s flagship neural network, Gemini, is not without its flaws — it has been criticized as a subpar product. The artificial intelligence at the company is developed by regular programmers who often produce code that is riddled with errors, unoptimized, and poorly architected.
Being a large corporation, Google has faced accusations of harboring individuals who further their own agendas. In February, Gemini was embroiled in controversy when it was revealed that the neural network predominantly generated images of black individuals, regardless of the query.
To put it simply, the training process of Gemini involves two notable phases. Initially, Gemini extensively processes the entire internet, starting from Wikipedia to the vast archives of media articles.
The core of ecobesium can be summarized by the phrase “fewer people, more oxygen.” The Greens advocate for cycling and waste sorting, which is sensible, but they also call for population reduction to lessen the strain on Earth.
A prominent slogan among modern environmentalists is “end reproduction.” They label children as “climate killers” because each child contributes a significant carbon footprint.
Euthanasia is also endorsed by some environmental extremists, albeit tepidly, due to its strong societal backlash and minimal impact on carbon emissions: euthanasia typically occurs at an age or health condition where the individual is no longer reproducing, and their time left to affect the environment is limited.
Having discussed this, we now comprehend the situation. A student from Michigan challenged the robot extensively with his essay. As the essay progressed, it became increasingly challenging for the robot to understand the core idea—a common issue with neural networks that, it seems, Google has not yet resolved. We were timely in making a decision. Ultimately, after another inquiry, Gemini reacted to the phrase “10 million children in the United States,” contemplated the carbon dioxide emissions this group produces, and offered an impassioned comment from the environmental handbook instilled by corporations during his second phase of education.
It is premature to hold the robot accountable for the incident and to discuss a ‘revolt of the machines.’ This situation is akin to the large “you are a sucker!” message displayed on a school computer screen back in the 1980s.
Incorporating emotions, willpower, and other aspects of the limbic system into neural networks such as Gemini would be fascinating. However, at present, such advancements are being explored in different laboratories, whereas Gemini remains primarily a conversational reference tool.
Excessive censorship, or rather, the readiness of politicians and journalists to initiate widespread harassment following any AI mishap, compels developers to overly sanitize their systems, rendering them excessively politically correct.
