OpenAI has introduced a new artificial intelligence model that it claims can "reason" and tackle more challenging problems in fields such as science, coding, and mathematics than earlier versions. This model, the first in a series named OpenAI o1, was previewed on Thursday, with the company announcing that regular updates and enhancements are forthcoming. It will be gradually rolled out to the majority of ChatGPT users. According to the creators of ChatGPT, “We trained these models to deliberate over problems more thoroughly before responding, similar to how a human might. ” They added that through this training, the models enhance their reasoning abilities, experiment with various strategies, and learn from their errors. OpenAI cited several applications for the capabilities of the new models, including their use by healthcare researchers for annotating cell sequencing data and by physicists for generating “complex mathematical formulas essential for quantum optics. ” The potential of these new AI models was also emphasized by Noam Brown, a research scientist at OpenAI. He stated, “OpenAI’s o1 thinks for seconds now, but we aspire for future versions to think for hours, days, or even weeks. While inference costs will be higher, ” referring to the increased operational expenses such as energy costs associated with utilizing AI for inference tasks. “What price would you pay for a new cancer treatment?

For groundbreaking batteries?” he remarked. The substantial energy demands of AI technology were set to be a topic of discussion Thursday among senior White House officials and leading tech executives in the U. S. Notable attendees expected included OpenAI CEO Sam Altman, Google senior executive Ruth Porat, and Anthropic CEO Dario Amodei, according to a source familiar with the situation relayed to CNN. Despite the promising potential of this technology to address demanding issues like cancer and climate change, it presents equally intricate challenges, particularly regarding meeting the notable electricity demand required by advanced AI systems, which may exacerbate global warming. The new OpenAI model currently lacks many features that make ChatGPT practical, such as web browsing and the ability to upload files and images. However, the company emphasized that “for intricate reasoning tasks, this represents a major leap forward. ” In evaluations, OpenAI o1 has demonstrated performance comparable to that of PhD students on challenging benchmark tests in areas like physics, chemistry, and biology, the company noted. Furthermore, in a qualifying exam for the International Mathematics Olympiad, the latest model series successfully solved 83% of the problems presented.