What is AGI? and what does it have to do with OpenAI?

This is potentially a game-changer in a lot of ways, including some ethical discussions about how it could end up affecting or replacing human jobs. (Representational Image)

Artificial General Intelligence (AGI) has been in the news a lot recently, thanks to the heap of trouble brewing over at OpenAI.

In the last few days, all hell had broken loose at the AI company, CEO's were changed, board members were replaced, and ,when all was said and done, Sam Altman found himself back as the CEO. It all apparently started because of disagreements about safety concerns about new breakthroughs that the company had made in AGI.

When you think of AI chatbots like Copilot (formerly Bing Chat), and ChatGPT, they are highly focused, specialised language models that have been trained to be good at a specific use case -- in this case, holding natural conversations with the user.

AGI, or Artificial General Intelligence, is more a general term, referring to artificial intelligence that has been trained to excel not just in one particular task, but to be good at nearly everything.

This also includes the ability to learn, and AGI models can teach themselves, potentially improving their capabilities even more over time. Rather than be focused on one specific task like writing greeting cards for example, an AGI can also pick out the design, the dimensions, the print quality and the paper used, all while explaining to the customer why they are paying Rs 50 extra for the fancy one.

This is potentially a game-changer in a lot of ways, including some ethical discussions about how it could end up affecting or replacing human jobs.

AGI isn't quite here yet, and experts are divided on when it actually will be reality, some say that it may take twenty years, while others believe it's merely a few years away.

Another interesting scenario to think about is whether it will pose an existential threat to human beings. Most experts seem to agree that while AGI, if it happens, will most likely achieve independence from humans in some form, it will not be an existential threat.

According to Reuters, several researchers within OpenAI sent the board of directors a letter warning them of a new discovery within AGI that could threaten humanity. This was before Sam Altman was ousted from the CEO's role.

The letter was apparently one of the factors in a long list of problems the board seemed to have with Altman, which ultimately led to his firing. Mira Murati, OpenAI CTO, and the company's interim CEO for one day, had mentioned the project codenamed "Q" to employees before the Altman saga.

Q, a.k.a. Q-Star, is supposed to be the company's supreme breakthrough in super-intelligence or AGI. The researchers were excited about the math capabilities of the model. This is because AI's today are good at predicting or writing, but solving mathematical equations where there is only one right answer is more difficult.

The letter the researchers sent to the board, flagged the safety concerns associated with a model that could potentially pose danger to humanity.