Former Employees of Top AI Firms Sign an Open Letter Warning about the Risks of AI

Former employees of top Silicon Valley firms such as OpenAI, Google’s DeepMind, and Anthropic have signed an open letter, sending out a stark warning about the risks of AI and how it could even lead to human extinction.

To clarify his stance on the issue, Neel also put up a post on X where he said that he only wants companies to guarantee that if there’s a concern with a certain AI project, the employees will be able to warn against it without repercussions i.e. whistleblowing freedom.

He further added that there’s no immediate threat that he wants to warn about, and that the letter is just a precautionary step for the future. Now, as much as I’d like to believe Neel on this, the content of the letter paints a different picture.

What Does the ‘Warning Letter’ Say?

The letter acknowledges the benefits AI advancement can bestow upon society, but it also recognizes the numerous downsides it tags along.

The following risks have been highlighted:

In short, everything we see in an apocalyptic sci-fi movie (such as Arrival and Blade Runner 2049) can come to life.

The letter also argued that the AI firms are not doing enough to mitigate these risks. Apparently, they have enough “financial incentive” to focus more on innovation and ignore the risks for now. Worrisome, to say the least.

This is in reference to the latest controversy at OpenAI, where employees were forced to choose between losing their vested equity or signing a non-disparagement agreement that would be forever binding on them.

The company later retracted this move, saying that it goes against its culture and what the company stands for, but the damage had already been done.

For example, in May pf this year, OpenAI disbanded the team that was responsible for researching the long-term risks of AI less than a year after it was formed.

However, it’s well worth noting that the company did form a new Safety & Security Committee last week. It will be headed by CEO Sam Altman.

Several high-level executives also left OpenAI recently, including co-founder Ilya Sutskever. While some were let go of with grace and sealed lips, others such as Jan Leike revealed that OpenAI has digressed from its original objectives and is no longer prioritizing safety.

OpenAI’s Response to This Letter

Addressing the above-mentioned letter, an OpenAI spokesperson said that the company understands the concerns surrounding AI and firmly believes that a healthy debate over this matter is crucial.

Furthermore, they said that OpenAI will continue to work with the government, industry experts, and communities around the world to develop AI safely and sustainably.

“We’re proud of our track record providing the most capable and safest AI systems and believe in our scientific approach to addressing risk.” – OpenAI

It was also pointed out that whatever new regulations have been imposed to control the AI industry have always been supported by OpenAI.

Quite recently, OpenAI disrupted 5 covert operations backed by China, Iran, Israel, and Russia that were trying to misuse AI-generated content and debug websites and bots to spread their malicious propaganda.

Speaking of giving employees the freedom to voice their concerns, OpenAI highlighted that it already has an anonymous hotline in place for its workers for this exact reason i.e. anyone can report their concerns regarding the company’s dealings without revealing their identity.

Despite what the company says about taking safety-oriented steps, we never completely know what’s actually going on. This is perhaps the most frightening part—that we will perhaps never know of a big misstep in AI development until it’s too late.

For example, although all of the above-mentioned companies have policies against using AI to create election-related misinformation, there’s evidence that OpenAI’s image-generation tools have been used to create misleading content.