Unlock Editor's Digest totally free
Roula Khalaf, editor of the FT, picks her favorite stories on this weekly newsletter.
OpenAI expects artificial intelligence assistants to “reach the mainstream” next yr as tech giants including Google and Apple vie to supply so-called AI agents to consumers.
AI agents that may solve and solve complex tasks for humans have grow to be the newest front within the battle between tech firms in search of to extend their revenue from the rapidly evolving technology.
“We have the desire to make it possible to interact with AI in all of the ways you interact with one other human,” said Kevin Weil, chief product officer at OpenAI.
“These more agent-based systems might be possible, and that's why I believe 2025 might be the yr that agent-based systems finally reach the mainstream,” he added.
At its developer day in San Francisco on Tuesday, OpenAI unveiled improved access to its latest model line called o1, which features improved reasoning in addition to the advanced language capabilities of GPT-4o. Developers can access this technology in real time, with the AI ​​in a position to understand voice commands and converse verbally in a live scenario just like a call.
The push to make AI agents available to the masses is a method OpenAI expects its technical advances to assist generate future profits as the corporate advances its plans to restructure as a for-profit company.
The fast-growing startup is aiming to shut a $6.5 billion funding round this week at a $150 billion valuation by convincing backers that it’s able to taking over its rivals to beat at key technological milestones and dominate the sector. Investors in discussions with the corporate in recent weeks included Microsoft, Nvidia, SoftBank and enterprise capital firms Thrive Capital and Tiger Global, based on people aware of the discussions.
Microsoft, Salesforce and Workday last month put agents at the middle of their AI plans, while Google and Meta also indicated it will be a magnet for them as they integrate their AI models into their products.
While AI-powered assistants have been in use for nearly a decade, these latest advances enable smoother and more natural language interactions and better levels of understanding because of the massive language models (LLMs) that power latest AI models.
Last yr, OpenAI released the “Assistant Application Programming Interface,” which goals to permit developers to create agents using its technology. However, the corporate said this was hampered attributable to limited capabilities of previous models.
Weil said the improved pondering and reasoning capabilities of OpenAI's latest models would manifest in its products, including ChatGPT, in addition to in startups and developers constructing products using its API. The company declined to comment on whether it immediately plans to construct its own AI agent.
One example shown in a bespoke demonstration of the tools was communicating with an AI system to assist source products that could be purchased locally, akin to strawberries. The AI ​​would then call the corporate to put an order of strawberries, taking the user's instructions regarding what number of and the way much they would love to spend.
OpenAI said that any use of such technology mustn’t hide the proven fact that it’s AI and never human, and that it will only be available to developers in six presets, slightly than developing latest voices.
“If we do it right, we'll get to a world where we will actually spend more time doing the things that matter and a little bit less time looking at our phones,” Weil said.
