SAN FRANCISCO (News21USA) — Less than a year after its meteoric rise, the company behind ChatGPT on Monday revealed the future it has in mind for its artificial intelligence technology by launching a new line of chatbot products that can be customized for a variety of tasks.
“Over time, you’ll just ask the computer what you need and it will do all these tasks for you,” OpenAI CEO Sam Altman told a crowd of more than 900 software developers and other attendees. It was the inaugural OpenAI developer conference, embracing a Silicon Valley tradition for technology showcases that Apple helped pioneer decades ago.
At the event held at a cavernous former Honda dealership in San Francisco, OpenAI’s hometown, the company unveiled a new version called GPT-4 Turbo that is “more capable” and can retrieve information about world and cultural events as recent as April 2023, unlike previous versions that couldn’t answer questions about anything that happened after 2021.
It also touted a new version of its AI model called GPT-4 with Vision, or GPT-4V, which allows the chatbot to analyze images. In a September research paper, the company showed how the tool could describe what’s in images to people who are blind or have low vision.
Altman said ChatGPT has more than 100 million weekly active users and 2 million developers, spread “completely by word of mouth.”
Altman also introduced a new line of products called GPT (emphasis on plural) that will allow users to create their own customized versions of ChatGPT for specific tasks.
Alyssa Hwang, a computer science researcher at the University of Pennsylvania who first got a glimpse of the GPT vision tool, said it was “very good at describing a lot of different types of images, no matter how complicated,” but also It needed some improvements.
For example, in trying to test its limits, Hwang added an image of a steak with a caption about chicken noodle soup, confusing the chatbot by describing the image as if it had something to do with chicken noodle soup.
“That could lead to some adversarial attacks,” Hwang said. “Imagine if you put offensive text or something like that on an image, you’ll end up getting something you don’t want.”
That’s partly why OpenAI has given researchers like Hwang early access to help discover flaws in its newest tools before their widespread release. Altman on Monday described the company’s approach as “gradual iterative implementation” that allows time to address security risks.
The road to OpenAI’s debut DevDay has been unusual. Founded as a nonprofit research institute in 2015, it catapulted to global fame just under a year ago with the launch of a chatbot that generated excitement, fear and a push for international safeguards to guide rapid advancement of AI.
The conference comes a week after President Joe Biden signed an executive order that will establish some of the first guardrails in the United States for artificial intelligence technology.
Using the Defense Production Act, the order requires AI developers likely to include OpenAI, its financial backer Microsoft, and competitors such as Google and Meta, to share information with the government about AI systems being built with “levels of such high performance” that they could pose serious safety risks.
The order was based on voluntary commitments set by the White House that leading AI developers made earlier this year.
There are also many expectations placed on the economic promise of the latest generation of generative artificial intelligence tools that can produce passages of text and images, sounds and other novel media in response to written or spoken prompts.
Altman was briefly joined on stage by Microsoft CEO Satya Nadella, who said to applause from the audience, “We love you guys.”
In his comments, Nadella emphasized Microsoft’s role as a business partner that uses its data centers to provide OpenAI with the computing power it needs to build more advanced models.
“I think we have the best partnership in technology. I’m excited that we can build AGI together,” Altman said, referencing his goal of building so-called artificial general intelligence that can perform as well (or even better) than humans on a wide variety of tasks.
While some commercial chatbots, including Microsoft’s Bing, are now built on OpenAI’s technology, there are a growing number of competitors, including Google’s Bard and Claude from another San Francisco-based startup, Anthropic, led by former OpenAI employees. OpenAI also faces competition from developers of so-called open source models who publish their code and other aspects of the system for free.
ChatGPT’s new competitor is Grok, which billionaire Tesla CEO Elon Musk unveiled over the weekend on his social media platform X, formerly known as Twitter. Musk, who helped start OpenAI before parting ways with the company, launched a new company called xAI this year to make his own mark on the pace of AI development.
Grok is only available to a limited set of early users, but promises to answer “spicy questions” that other chatbots reject due to security measures intended to prevent offensive responses.
When asked by a reporter for comment on the timing of Grok’s release, Altman said, “Elon is going to be Elon.”
Goldman Sachs projected last month that generative AI could boost labor productivity and lead to a long-term increase of 10% to 15% in global gross domestic product — the economy’s total output of goods and services.
Altman described a future of AI agents that could help people with various tasks at work or at home.
“We know that people want AI that is smarter, more personal, more customizable and can do more for you,” he said.