Anthropic Thinks Its Own Success Matters In Making AI Safe

Anthropic has been around for the past five years to warn the world about how advanced weaponry can lead to mass destruction, destabilization, and many other dangers. But at the same time, it has become one of the most powerful forces pushing the potential of AI forward. The company is now one of the leading developers and distributors of AI-powered models for law enforcement customers such as the US military. Recently it was appreciated about $1 trillion.

At first glance, Anthropic’s powerful messages and actions seem contradictory.

But inside the company, most people don’t see the contradiction. To understand why, you must first understand that Anthropic operates based on two core beliefs. The first is that artificial intelligence is the most transformative technology in human history, and its arrival is inevitable. The real question is whether it brings disaster or incredible prosperity.

The second is that Anthropic believes the world will be better off if it stays on the cutting edge of AI competition, according to several former employees who spoke to WIRED on condition of anonymity. Internally, leaders and employees at the company often refer to themselves as “the good guys,” meaning those in charge of AI technology, two of the sources said. The company sees accumulated power – whether it’s capital, computing, research skills, or political influence – not as an end in itself, but as the cost of achieving it. work: “ensuring a better world through AI revolution.”

Helen Toner, director of Georgetown’s Center for Security and Emerging Technology and former OpenAI board member, uses it illustration expressing the Anthropic worldview. He compares the powerful AI to a forest full of magic and dangerous monsters. All the people in the nearby villages are running, attracted by the treasure. In his words, Anthropic wants to go deeper into the jungle than anyone else while investing heavily in dealing with these monsters, that is, taking the advantages of AI where it has its disadvantages.

“The difference with Anthropic is, ‘People are going to the forest, we have to do it first.’ This is their obvious strategy: to build the most expensive AI to be the biggest player at the table who can talk about what advanced AI systems should look like, what risks they pose, and insist on adequate security,” Toner tells me. “He’s very blunt about it. It’s a very strange way that people have a hard time hearing.”

Anthropic CEO Dario Amodei explained this clearly in a discussion with its founders posted on the company’s job site: “You have to find a way to really be a competitor, really to lead companies sometimes, but they manage to do things carefully,” he says. If you can do that, the gravity you exert is enormous.”

Anthropic was it was launched in 2021 is a group of former OpenAI employees who left after losing faith in the ability of the company’s leadership — especially CEO Sam Altman — to seamlessly bring AI to the world. Those ideas still shape the company to this day. Two of the former employees I spoke with say that, in internal discussions, Anthropic executives often point to Altman and OpenAI—and, to a lesser extent, Meta and Elon Musk’s xAI—as cautionary examples that help explain Anthropic’s responsible thinking.

In general, Anthropic is like any other Silicon Valley company. Many startups sell themselves as David versus the old Goliath, an established industry that wants to disrupt. Google, Facebook, and Apple were all founded on good ideas, which were later undermined or abandoned as they got bigger, bigger, and more popular.