The political world did not seem to know how seriously these warnings should be heeded. Asked if AI is dangerous, President Biden said on Tuesday: “That remains to be seen. May be.”

Many in the insular AI sector of Silicon Valley are familiar with the dystopian visions where a small group of strange but powerful subcultures have clashed in recent months. One sect believes that AI can kill us all, another argues that this technology will allow humanity to flourish if used correctly. Others are suggesting a six-month pause suggested by Musk, who will reportedly launch his own artificial intelligence laboratorywas designed to help him catch up.

Sub-groups can be quite fluid, even if they seem to be contradictory, and insiders sometimes disagree on the basic definitions.

But these once-marginal worldviews can become key to discussions about AI. Here is a quick guide to deciphering the ideologies (and financial incentives) behind the facts:

Argument: The phrase “AI safety” used to refer to practical issues, like making sure self-driving cars don’t crash. In recent years, the term, which is sometimes used interchangeably with “AI compliance”, has also been adopted to describe a new area of ​​research aimed at making AI systems obey the intentions of their programmers and prevent AI’s quest for power that can only harm humans. in order to avoid shutdown.

Many have ties to communities such as effective altruism, a philosophical movement to maximize the good in the world. EA famously started by prioritizing causes like global poverty, but focused on concerns about the risk associated with advanced AI. Online forums like Lesswrong.com or the AI ​​Alignment Forum have heated debates about these issues.

Some adherents also have a philosophy called longevity, which aims to maximize the good over millions of years. They cite a thought experiment from Nick Bostrom’s Superintelligence that suggests that safe superhuman AI could allow humanity to colonize the stars and create trillions of future humans. Creating a safe artificial intelligence is critical to keeping these possible lives safe.

Who is behind this: In recent years, EA-affiliated donors such as Open Philanthropy, a foundation founded by Facebook co-founder Dustin Moskowitz and former hedge fund Holden Karnofsky, have helped create a number of centers, research labs, and community-building efforts focused on AI and AI security. alignment. FTX Future Fund, founded by crypto chief executive Sam Bankman-Freed, was another big player until the firm went bankrupt after Bankman-Freed and other executives were accused of fraud.

What influence do they have?: Some work in leading AI labs such as OpenAI, DeepMind, and Anthropic, where this mindset has led to some useful ways to make AI safer for users.. A close-knit network of organizations is conducting research and research that can be disseminated more widely, including 2022 survey it showed that 10 percent of machine learning researchers say AI could end humanity.

AI Impacts, which conducted the study, received support from four different EA-affiliated organizations, including the Future of Life Institute, which hosted Musk’s open letter and received the largest donation from Musk. Center for Humane Technology co-founder Tristan Harris, who once campaigned on the dangers of social media and has now focused on AI, cited the study extensively.

Argument: It’s not that this group doesn’t care about security. They are just very excited about creating software that achieves artificial general intelligence, or AGI, is a term for an AI that is as smart and capable as a human. Some of them are promising tools like GPT-4, which OpenAI says have developed skills like writing and answering in foreign languages ​​without any instruction, which means they are on their way to AGI. Experts explain that GPT-4 has developed these capabilities by ingesting huge amounts of data. most say that these tools do not have a human understanding of the meaning of the text.

Who is behind this?: Two leading artificial intelligence labs mentioned the creation of AAI in their mission statements: OpenAI, founded in 2015, and DeepMind, a research lab founded in 2010 and acquired by Google in 2014. wealthy technology investors interested in the outer limits of AI. According to Cade Metz’s The Genius Makers, Peter Thiel donated $1.6 million to Yudkowsky’s non-profit artificial intelligence organization, and Yudkowsky introduced Thiel to DeepMind. Musk invested in DeepMind and introduced the company to Google co-founder Larry Page. Musk introduced the concept of AGI to other OpenAI co-founders such as CEO Sam Altman.

What influence do they have?: OpenAI’s dominance of the market has thrown open the Overton window. The CEOs of the world’s most valuable companies, including Microsoft CEO Satya Nadella and Google CEO Sundar Pichai, are now being asked and discussed AGI in interview. Bill Gates blog about it. “Because the benefits of AGI are so great, we don’t think it’s possible or desirable for society to permanently stop its development,” Altman. wrote in February.

Argument: Although doomers share a range of beliefs and often visit the same online forums as people in the AI ​​security world, this crowd came to a conclusion that if you connect a powerful enough AI, it will destroy human life.

Who is behind this?: Yudkowsky led the warning voice of this doomsday scenario. He is also the author of the popular Harry Potter and the Methods of Rationality fanfiction series, which is a starting point for many young people in these online fields and ideas related to artificial intelligence.

His non-profit organization MIRI received support $1.6 million in donations in its early years from technology investor Thiel, who has since distanced himself from the group’s views. EA-affiliated organization Open Philanthropy donated about $14.8 million under five grants from 2016 to 2020. More recently, MIRI has received funds from crypto nouveau riches, including Ethereum co-founder Vitalik Buterin.

What influence do they have?: Although some in this world consider Yudkowsky’s theories prophetic, his writings have also been criticized for not being applicable to modern machine learning. However, his views on AI have influenced more vocal voices on these topics, such as renowned computer scientist Stuart Russell signing an open letter.

In recent months, Altman and other raised the authority of Yudkovsky. Altman recently tweeted that “maybe at some point [Yudkowsky] deserves the Nobel Peace Prize” for accelerating AGI, and later also tweeted a photo of the two of them at a party hosted by OpenAI.

Argument: For years, ethicists have been warning about problems with larger AI models, including results that are biased by race and gender, the explosion of synthetic media that could damage the information ecosystem, and the impact of AI that sounds deceptively human. . Many argue that the apocalypse narrative exaggerates the capabilities of AI. helping companies advance technology as part of science fiction fantasy.

Some in this camp argue that the technology is not inevitable and can be created without harming vulnerable communities. Technology-focused criticism can ignore human decisions, allowing companies to avoid liability for poor medical advice or privacy violations by their models.

Who is behind this?: co-authors visionary research work warning about the dangers of large language models, including Timnit Gebru, former co-lead of Google’s Ethical AI Group and founder of the Distributed AI Research Institute, are often cited as leading voices. Critical research demonstrating the failures of this type of AI, as well as ways to mitigate the problems, is “often conducted by scientists of color — many of them black women” and underfunded by junior scientists, researchers Ababa Birhain and Deborah Raji. wrote in an article for Wired in December.

What influence do they have?: In the midst of the AI ​​boom, tech firms like Microsoft, Twitch and Twitter are laying off their AI ethics teams. Bootmakers and the public listened to politics.

Former White House politician Suresh Venkatasubramanian, who helped draft the AI ​​Bill of Rights, said VentureBeat claims that the recent exaggerated claims about ChatGPT’s capabilities were part of an “organized fear-mongering campaign” around generative AI that distracted from stopping work on real AI problems. Gebru spoke before European Parliament about the need for a slow movement of AI, slowing down the pace of development of the industry, so that the safety of society is in the first place.