OpenAI, a pioneering AI company backed by Microsoft, has introduced a comprehensive safety framework for its most cutting-edge models in a decisive move aimed at addressing the growing concerns surrounding artificial intelligence safety.
The framework, disclosed via a plan released on its official website on Monday, includes a notable provision allowing the company’s board to reverse safety-related decisions, Reuters news report said.
OpenAI’s ChatGPT AI tool is used by 100 million people on a weekly basis, OpenAI CEO Sam Altman said recently at the developer forum. Since releasing its ChatGPT and Whisper models via API in March, the company also now boasts over two million developers, including over 92 percent of Fortune 500 companies.
The development assumes significant importance because the deployment of AI will be facing regulatory challenges in several markets.
Under the outlined strategy, OpenAI will limit the deployment of its latest AI technology exclusively to areas vetted as safe, particularly focusing on domains like cybersecurity and nuclear threat mitigation.
To bolster safety protocols, the company is establishing an advisory group tasked with scrutinizing safety reports and forwarding them to both company executives and the board. While the executives will be responsible for making determinations, the board possesses the authority to overturn these decisions, emphasizing a multi-tiered approach to ensuring AI safety.
Ever since the debut of ChatGPT a year ago, concerns regarding the potential hazards of AI have been paramount among AI researchers and the wider public. While generative AI technology has captivated users with its proficiency in crafting poetry and essays, it has concurrently raised apprehensions due to its capacity to disseminate misinformation and manipulate human behavior.
The unease surrounding AI’s implications reached a crescendo in April when a coalition of AI industry leaders and experts penned an open letter advocating for a six-month hiatus in developing systems surpassing the capabilities of OpenAI’s GPT-4, citing inherent risks to society.
Subsequent to this, a Reuters/Ipsos poll conducted in May disclosed that more than two-thirds of Americans harbor concerns regarding the potential adverse impacts of AI, with 61 percent expressing apprehensions about its potential threat to civilization.
