Google parent Alphabet on Tuesday showed how it is leveraging artificial intelligence across its business, including a beefed-up Gemini chatbot and search improvements, as it faces growing competition from OpenAI and other competitors.
At its annual I/O developer event in Mountain View, Calif., CEO Sundar Pichai said the company is rolling out AI previews to all users in the United States this week, after a long public testing period since last year.
The new AI features unveiled Tuesday will help investors gauge Alphabet’s progress as it battles Microsoft, OpenAI and other rivals to dominate the emerging technology.
Alphabet shares climbed during the product presentation and were last up about 1% at $172.50 Tuesday afternoon.
“We are in the very early days of AI platforms,” Pichai said.
Google has announced improvements to its Gemini Pro 1.5 model, capable of making sense of massive amounts of data.
On Tuesday, Google announced it would double that amount to 2 million tokens, meaning the AI could potentially answer questions when given thousands of pages of text or more than one hour of video to ingest.
The Pro model – starting with prompt sizes up to 1 million tokens or data – will also be available to subscribers of Google’s Gemini Advanced service.
In another sign of fierce competition between OpenAI and Google, the online search leader teased Veo, an AI model it claims is the most powerful yet for creating videos on a simple text command.
Google released earlier video generation technology in January, only to be eclipsed a few weeks later by OpenAI’s Sora.
The creator of ChatGPT has been promoting his filmmaking software to Hollywood executives, captivating and worrying the creative industry.
Google said filmmaker Donald Glover experimented with its AI.
The company also previewed a new text-to-image conversion model, Imagen 3, and touted collaborations from other artists.
The company announced a scaled-down version of Gemini called 1.5 Flash, which aims to reduce the cost of deploying AI and speed up responses.
As the highest performance version, Flash can absorb large amounts of data while being optimized for chat applications, video and image captioning.
AI Overviews uses generative AI to synthesize information and answer more complex queries for which there are no simple answers on the web.
Alphabet’s AI unit, Google DeepMind, has been working to create technology that can perform everyday tasks for consumers.
The first results were manifested in the Astra project, a tool capable of using a smartphone camera and drawing conclusions about the world around it.
In a demo video shown during Google I/O, a user deployed it to identify a speaker and locate the glasses they had left in another part of the room.
Microsoft-backed OpenAI on Monday introduced a new AI model called GPT-4o, which allows ChatGPT to respond by voice in real time and be interrupted – two characteristics of realistic voice conversations that AI voice assistants like Google Assistant found it difficult.
New York Post
