OpenAI is embarking on a new chapter in its journey toward artificial general intelligence (AGI). The company has started training its next major AI model, which promises to elevate AI capabilities significantly beyond what GPT-4 offers. This ambitious project aims to push the boundaries of AI technology, taking OpenAI a step closer to achieving AGI.
To oversee the development and ensure the safety of this new model, OpenAI has established a dedicated Safety and Security Committee. This committee is led by CEO Sam Altman, Chairman Bret Taylor, and board members Adam D'Angelo and Nicole Seligman, along with technical and policy experts from within the company. Their immediate task is to evaluate and enhance OpenAI's safety protocols over the next 90 days, after which they will present their findings and recommendations to the board.
OpenAI acknowledges the critical nature of this moment in AI development and has called for a robust debate on the subject. "While we are proud to build and release models that are industry-leading in both capabilities and safety, we welcome a rigorous debate at this important moment," the company stated. This proactive approach comes in the wake of the departures of key figures like co-founder Ilya Sutskever and Jan Leike, who previously led OpenAI's long-term safety efforts.
The forthcoming model, anticipated to be significantly larger than GPT-4, has been likened to a "giant whale" compared to GPT-4's "Orca" by Microsoft CTO Kevin Scott. Although specific details about the new model remain under wraps, it is clear that it aims to bring substantial advancements in AI capabilities.
Interestingly, while OpenAI is accelerating its AI development, there seems to be a slight shift in its communication regarding its ultimate goals. Previously, CEO Sam Altman had expressed an aim to build a "superintelligence" far surpassing human intelligence. However, Anna Makanju, OpenAI’s vice-president of global affairs, recently clarified that the mission is to build AGI capable of performing cognitive tasks akin to those humans can accomplish today. "Our mission is to build AGI; I would not say our mission is to build superintelligence," Makanju stated. She emphasized that superintelligence, while a subject of study, is not the company's primary mission.
OpenAI's spokesperson, Liz Bourgeois, reinforced this position, stating that the mission is to create AGI that benefits humanity. She noted that while superintelligence is studied, it is generally viewed as more advanced than AGI. This clarification aims to dispel any perceived conflict in the company's goals.
As OpenAI moves forward, it is also addressing the need for responsible AI development, especially in light of recent staff departures and increasing competition from tech giants like Google's Gemini and Elon Musk's xAI. The new safety and security committee, led by Altman and other board directors, is a critical component of this effort, ensuring that the next-generation AI model is developed with the highest standards of safety and security in mind.
The training of this new model involves an extensive process where the AI learns from vast amounts of data, improving its performance through rigorous validation and testing phases. Given the complexity and scale of this endeavour, it may take several months before the new model is ready for deployment in products or applications.
OpenAI's pursuit of AI advancement, while prioritizing safety and ethics, represents a significant milestone in the development of artificial intelligence. By tackling these challenges head-on, the company continues to be a key player in shaping the future of AI.