OpenAI's Sam Altman calls AI systems like GPT-4 'safe enough to use'

Join We In to come back has new York on June 5th has collaborate with executive leaders In exploring complete methods For audit AI models concerning bias, performance, And ethics compliance through miscellaneous organizations. Find out how You can attend here.

OpenAI Chief Executive Sat Altmann briefly appeared has This years Microsoft Build has talk with from Microsoft Chief Technology Officer Kevin Scott. THE Hourly was notable because he came less that 24 hours After actress Scarlett Johansson accused OpenAI of using A voice "strangely similar" has hers For It is GPT-4o chatbot.

Altmann did not directly address THE dispute, neither did Microsoft, OpenAI the biggest partner. However, he did share that he was surprised by how a lot he appreciated voice fashion In GPT-4o. Regardless, OpenAI chief spent almost ten minutes giving A inspiring message about how developers are key has THE future of artificial intelligence innovation.

“Developers to have has been such A heart part of what is this has been event In THE last year And A half," he Remarks. "There are millions of people WHO to have has been on THE platform. What people are TO DO East totally amazing. THE speed of adoption And Talent And to understand out What has build with all of This on What has Really not has been very long time... when We put GPT-3 In THE API, a few people thought he was cool, but he was zero income. And seeing What people to have do with GPT-4 And NOW what is this event with GPT-4o… it’s enough outstanding. »

" We have Never seen A technology get adopted SO quickly In such A significant path," he proclaims.

THE AI Impact Tour: THE AI Verification

Join We as We back has new York on June 5th has to commit with high executive leaders, rummage In strategies For audit AI models has ensure justice, optimal performance, And ethics compliance through miscellaneous organizations. Secure your presence For This exclusive invite only event.

When request What We should to wait for on THE following little month, Altmann said models will continue has earn In intelligence. "If You think about What arrived Since GPT-3.5 has GPT-4, he just obtained smarter… that obtained A little more robust, a lot safer because THE model obtained more intelligent And put a lot more work In building THE security tools around he. He obtained more useful. »

Later, he would be talk more about AI security, A area her business East often partner with, especially following A cut by advice members last November, THE departure of key executives And THE dismantling of It is superalignment team. He actions that When departure out, OpenAI spent A plot of time trying has get It is AI has A acceptable level of security. With GPT-4, It is assumed has be on enough. "It is far Since Perfect [And] We to have more work has TO DO," Altmann States. "But he East in general considered robust enough And on enough For A wide variety of uses. And that took A huge Rising of work through both teams And ...