What is Google’s ‘Jarvis’ AI and why is it different?

Google is said to be developing an AI system called Project Jarvis which can use computer software by itself - booking flights and buying products for users.

LOS ANGELES, USA - SEP 28, 2015: Iron Man in the  Madame Tussauds Hollywood wax museum. Marie Tussaud was born as Marie Grosholtz in 1761
The software is named after the Jarvis AI in Iron Man, pictured in Madam Tussauds Hollywood (Alamy)

Google is said to be developing an artificial intelligence system called Project Jarvis which can use computer software by itself - booking flights and buying products for users.

Project Jarvis, named after the AI system used by Tony Stark in Iron Man, is an ‘AI agent’, a new type of artificial intelligence which can 'take over' people's computers and use software such as web browsers to fulfil commands. Jarvis is capable of using a browser and taking actions such as buying plane tickets, for example, according to The Information.

Project Jarvis is expected to launch as early as December 2024, although it will initially only be available to a small group of testers. Other AI companies such as OpenAI and Anthropic are also developing ‘agentic AI’, with analysts predicting that the market could be worth $65 billion (£50 billion) by 2030.

Agentic AI (or AI agents) refers to AI software which is built to act autonomously from humans to fulfill certain goals.

ADVERTISEMENT

So, for example, you might say, “Find me the cheapest flight to Malta on December 4 and book me two seats on it,” and the AI agent would use browser software to do so for you.

CHINA - 2024/06/21: In this photo illustration, the Gemini, formerly known as Bard, is a generative artificial intelligence chatbot developed by Google (Alphabet), logo seen displayed on a smartphone with an economic stock exchange index graph in the background. (Photo Illustration by Budrul Chukrut/SOPA Images/LightRocket via Getty Images)
The software is expected to launch in preview alongside Google's new Gemini 2.0 AI (Photo Illustration by Budrul Chukrut/SOPA Images/LightRocket via Getty Images)

Bill Gates believes AI agents will completely change the way people interact with computers: rather than opening apps, people will simply give plain-language commands to a smartphone or PC and agents will do tasks for them.

Gates has been imagining AI agents for 30 years, having written about them in his 1995 book, ‘The Road Ahead’.

In a recent blog he said that current advancements in AI are turning agents into a rel possibility.

He wrote, in his GatesNotes blog, “You won’t have to use different apps for different tasks. You’ll simply tell your device, in everyday language, what you want to do.

“In the near future, anyone who’s online will be able to have a personal assistant powered by artificial intelligence that’s far beyond today’s technology.

Project Jarvis is expected to work within a browser (Google Chrome) and could automate ‘everyday’ tasks such as buying products, according to a report by The Information.

ADVERTISEMENT

The move follows the announcement of similar software by rival AI company Anthropic, which showed off AI software which can move the cursor, type text and press buttons on websites to fulfil commands from users.

The AI works by taking screenshots of what is on screen, interpreting those screenshots and taking actions such as filling in a search field or pressing a button.

At present, Jarvis is still relatively slow, the Information said.

The early preview of Jarvis will coincide with the launch of Google's powerful new Gemini 2.0 model.

Google’s Sundar Pichai describes AI agents as being “able to think multiple steps ahead, work across software and systems, all to get something done on your behalf, and, most importantly, under your supervision.”

For example, Google has shown off an agent which can automate the returns process, going through email to find a receipt and order number, and filling out the form by itself in response to a simple command.