What would you throw at OpenAI CEO Sam Altman given a chance to ask him anything? AGI risks? Its comparison to nuclear bombs? Misuse of data by OpenAI? Security threat? Privacy? Nah, how about love? Of all the questions posed by industry experts, government officials, and entrepreneurs in an ET interview with Altman in Delhi, only a few made OpenAI CEO ponder. The rest were a missed opportunity of sorts.
Wasted Opportunity?
On his current world tour, Altman is seeking to spread the word for ChatGPT as the company continues to expand and create its presence everywhere. The event was a golden opportunity to explore about OpenAI and pose questions surrounding the multiple announcements made by the company over the past few weeks. However, a majority of these questions revolved around generic AI information, such as risks of AI, existential crisis, work culture, and creative fulfilment — where ChatGPT couldn’t fit in.
Some of the questions could be called rhetorical, at best. Cred founder Kunal Shah’s question on what Altman learned about humans after ‘doing’ AI, for instance, received a meandering explanation of how Altman thinks “something strange” and “very important” is going on with humans which he hopes to preserve. There was another ill-fitting question by Prasoon Joshi on ‘creativity’ and how an individual can lose creative satisfaction with these systems, which again had nothing to do with OpenAI or Altman.
Source: Youtube
Déjà Vu
Sam Altman’s responses rang a few bells at various points, sounding too similar to the AI Senate hearing that happened a few weeks ago. It was either the questions that steered him to respond in a similar fashion, or that the answers, that were too familiar. Claims of him not having equity, wanting regulations to build and deploy responsible AGI, and the very famous “future systems can cure all diseases or help us address climate change” was like a well-scripted deal.
Announcing 1 million grant offers each for democratising AI and forming cybersecurity programs, the company is now inclining towards being a responsible company, working for the people and caring about the world by deflecting any AGI risk.
When asked what’s next for OpenAI, Altman was pretty nonchalant and responded with the same saga of how the model will get smarter, and multimodal with reduced hallucinations — something that the company is already claiming with their last update on process supervision training.
A Few Bouncers
Not everything was lost, though. A few questions touched upon the pressing issues of AI regulations and the evaluation of AGI. Srivatsa Krishna, IAS, asked about the company’s strategic move to ask the government for regulations, knowing how hard it is to bring collective action among nations. The question was the one that had crossed all our minds when Altman first mentioned about wanting an AI regulation in the Congress hearing. The response was sort of sour with Altman calling the question a ‘cynical take’ on things and defending the ‘existential risk’ as a need and saying that if governments are not able to act together, it will be left to the companies to cooperate.
A topic on AGI on determining the proximity to it was something that Fractal co-founder Srikanth Velamakanni brought up. He asked about the tests to know how close we are getting to AGI? Altman said that GPT-4 “is not close to AGI” and some of the evaluation parameters will be around the ability of the model to figure things on its own and discover and learn to solve problems that it has never seen before. While Altman believes they are nowhere close to AGI, one of the first few things he said at the start of the interview was that “OpenAI is all about the quest for AGI”.
When the tech experts watching the interview wished for questions that would nudged Altman to reveal the obscure side of OpenAI, came a bizarre one posed by one of them. Reeking of his obsession with the possibility of forming a relationship with a robot, he went: “With AI, can you remove what you find irksome about your lover and program a robot in order to get the perfect lover?” Altman was not only baffled, he sheepishly responded, “I hope we don’t fall in love with robots — that will be deeply depressing!”