OpenAI's chief technology officer, Mira Murati, believes that government regulators should be “actively involved” in setting safety standards for deploying advanced AI models such as ChatGPT.
She also argues that the proposed six-month moratorium on development is not the right approach to building more secure systems, and that the industry is not yet close to achieving artificial general intelligence (AGI) a hypothetical threshold of intelligence at which AI can perform any task. Tasks requiring intelligence, including human-level cognition. Her comments were made in an interview with The Associated Press published April 24. When asked about the safety precautions OpenAI took before rolling out GPT-4, Murati explained that the comp any took a slow approach to training, not only to curb the machine's penchant for bad behavior, but also to find out what's wrong with it. Any downstream issues related to class changes:
"You have to be very careful because you can create some other imbalance. You have to keep auditing so you have to adjust it again, and every time you intervene, you have to be very careful to see what else is being It's messy." In the wake of GPT-4, experts worried about the unknowns of AI's future, calling for interventions ranging from increased government regulation to a six-month moratorium on global AI development.
The Latter Proposal Has Been Followed and Support by Well-Known Figures in the Field of Artificial Intelligence Such As Elon Musk, Gary Marcus and Eliezer Yud. KOWSKI, WHILE BILL GATES, YANN Lecun and Wu Enda, ETC. MANY CELEBRITIES HAVE COME OUT AGainst It. For her part, Murati expressed support for the idea of increased government involvement, noting, “These systems should be regulated.” She continued: “At OpenAI, we've been working with governments and regulators and other organizations that are developing these systems. Dialogue, at least at the corporate level, to agree on some level of standards."
On the issue of stalled development, however, Muratti struck a more critical tone:
"Some of the statements in the letter are completely inconsistent with the development of GPT-4 or GPT-5. We are not training GPT-5. We do not have any plans to do so in the next six months. And we are not rushing to launch GPT-4. In fact, it took us six months to fully focus on the secure development and deployment of GPT-4." Responding to whether there is currently a "path between GPT-4 and things like AGI," Murati told The Associated Press, "We're nowhere near having a safe, reliable, consistent AGI system."
For those who think GPT-4 is close to AGI, this may be bad news. The company's current focus on security, and the fact that, according to Murati, it hasn't even trained GPT-5 yet, are strong indications that the coveted general intelligence discovery remains out of reach for now.
The company is increasingly focused on regulation as the trend toward government scrutiny grows. OpenAI recently banned its GPT product in Italy and faces an April 30 deadline to comply with local and EU regulations in Ireland a deadline experts say will be difficult to meet.
Such a ban could have serious repercussions for the European cryptocurrency space, as more people tend to adopt advanced cryptocurrency trading bots built on top of applications using the GPT API. If OpenAI and companies building similar products find themselves unable to operate legally in Europe, traders using the technology may be forced to look elsewhere.
