© IE Online Media Services Pvt Ltd
Tags:
Journalism of Courage
Artificial Intelligence is today ubiquitous with many industries deploying them for a variety of use cases. Today, big tech companies are scrambling to be the leader in AI. However, OpenAI has carved a formidable spot for itself with the launch of ChatGPT in November 2022. The San Francisco-based AI company has been consistently working to offer newer upgrades to its large language model, the GPT series.
Sam Altman, the CEO of the company, is among the most talked-about leaders in AI. With the plethora of use cases and ever-increasing loyalists of ChatGPT, Altman is often confronted with questions about what will be his company’s next big inventions – GPT-5 or AGI? AGI stands for Artificial General Intelligence and is perceived as the pinnacle of human creation, that is, whenever it comes to fruition. AGI is the representation of generalised human cognitive abilities in software. It is essentially a system that can perform any task that a human being is capable of. But while the prospects are many, so are the concerns.
Sam Altman and Open AI CTO Mira Murati recently sat down for an interview with The Wall Street Journal. In the nearly hour-long interaction, Altman and Murati shared about OpenAI’s ambitious vision when it came to AGI.
“They will produce a lot of productivity and economic value. We are talking about one system that can generalise across a lot of digital domains of human work,” Murati said.
On the other hand, Altman said that there are two things which, according to him, will matter the most over the next decade or in a few decades from now to improving the human condition. These are AGI and affordable, abundant energy. He said that AGI, characterised by highly advanced, comprehensive intelligence, will be transformative. The CEO feels that it will likely be humanity’s most potent tool, promising unprecedented problem-solving capabilities and creative expression.
“We’ll be able to express ourselves in new creative ways. We will make incredible things for each other, for ourselves, for the world, for this unfolding human story,” said Altman. Despite the inevitable challenges that come along with changes of such magnitude, Sam said that he sees a future where AGI’s benefits are so manifest that current reservations will seem unfounded. He envisions a society enriched by remarkable innovations, contributing to humanity’s ongoing narrative.
In response to a question inquiring about the expected timeline of AGI, Altman said that he believed it could happen in the next ten years. Murati added that she offered a definition, and she feels that discussions about intelligence often delve into its depth, consciousness, whether it is sentient, and other aspects. Suggesting how definitions change over time, Murati said that humanity has progressed from making machines that could excel at chess to the GPT series. She added that the concept continues to develop, pushing boundaries of how we define intelligence.
According to Altman, along with the changing perceptions, the shifting goalposts in terms of what is AI is compelling OpenAI to work harder.
When asked why he referred to AGI as median human, Altman said that for a long time, AI systems have been inferior to human performance in numerous tasks. “For instance, in areas where I possess significant expertise, I might observe that GPT-4 or even GPT-5 performs unsatisfactorily. However, in tasks where my skills are average compared to the global population, these systems might exhibit impressive capabilities.”
Altman said that he was implying that in any specialised field, human experts can achieve new heights, and doing the same may take AI considerably longer to replicate. However, when it comes to more general tasks or those areas where an individual’s skill level is not particularly high, AI’s ‘average’ performance can be quite effective. So, in tasks that I’m not skilled at, future iterations of these systems might provide substantial assistance.”
Meanwhile, when asked about when GPT-5 will arrive, Murati deflected by saying, “We are not there yet. It’s kind of a need-to-know basis. I will let you know.” While there has been no definitive timeline known about the arrival of AGI, tech companies seem to be working towards advancing their AI models. However, both Altman and Murati unanimously said that GPT-3, GPT-4, and GPT-4.5 all are steps towards AGI.
