Former OpenAI Experts Launch SSI for Secure AI Development

Safe Superintelligence, Inc. – Ensuring AI Aligns with Human Values

Artificial intelligence has made amazing progress in recent years, with language models like ChatGPT showing abilities that used to seem like science fiction. However, as AI systems become more advanced, worries about their safety and possible dangers have also increased. This concern is now the main focus of a new startup created by AI experts who used to work at OpenAI, one of the top AI research companies.

Safe Superintelligence, Inc. (SSI) was launched on June 19th by Ilya Sutskever, the former chief scientist at OpenAI, along with Daniel Levy, a former OpenAI engineer, and Daniel Gross, an investor and ex-partner at the prestigious startup accelerator Y Combinator. Their mission? To develop artificial intelligence that is not only extraordinarily capable but also provably safe and aligned with human values.

The founders stated their "singular focus" will be advancing AI safety and capabilities hand-in-hand, insulated from short-term business pressures. With offices in Palo Alto and Tel Aviv, SSI aims to attract top engineers and researchers dedicated to tackling this monumental challenge.

Shifts in Superalignment Team Impact OpenAI

Sutskever and Levy left OpenAI in May due to disagreements and changes within the company following Sam Altman's temporary dismissal as CEO. They were part of the "Superalignment" team formed in 2022 to tackle the challenge of managing superintelligent AI systems much smarter than humans, known as artificial general intelligence (AGI).

When Superalignment leads, including Jan Leike who moved to Anthropic, left OpenAI, the special safety team was disbanded. Despite this, OpenAI stays focused on developing AI responsibly through different efforts.

AI Safety Concerns Spark Industry Debate

Many influential people are worried about the future of super-intelligent AI. Vitalik Buterin, who helped create Ethereum, thinks that AGI (Artificial General Intelligence) is risky, but he believes that the bigger dangers come from companies being too ambitious and the military using AI for their purposes. 

Elon Musk and Steve Wozniak from Apple, along with more than 2,600 others, have signed a letter asking for a six-month break in training advanced AI. They want this break so that we can carefully think about and understand the risks involved.

Yet the race among tech giants and startups to push AI boundaries continues rapidly. While companies like Google, Microsoft, and Anthropic tout responsible development, critics argue profit motives could compromise safety considerations for transformative but unpredictable technology.

SSI Leads AI Safety for Human Welfare

This situation shows how important it is to have organizations like SSI that are fully committed to ensuring AI safety in the lead. As Sutskever and Levy pointed out, being free from commercial distractions allows these organizations to focus completely on making sure highly advanced AI systems align with human values and societal welfare.

Of course, tackling the big challenge of AI safety is a huge technical obstacle. Just figuring out clear goals and measures is a tricky philosophical issue that hasn't been settled yet. However, the SSI founders strongly think that the risks are so serious that AI safety should be the most important thing as the field moves ahead.

While excitement builds over AI's potential to help humanity, a growing group also warns we must remain watchful about the risks. As Musk said, "We're simply the least intelligent species there has ever been." The launch of SSI represents a high-profile initiative to ensure we don't create something smarter than us that leads to unexpected - and potentially disastrous - consequences.

Also read - Digital Rial CBDC Pilot Program Goes Live on Kish Island