What you need to know
- Google is reportedly working on “Project Jarvis,” an AI agent designed to automate simple tasks while browsing the web in Chrome.
- Alleged sources claim Jarvis works by taking screenshots of your display to “interpret” the next best course of action — and does it.
- Jarvis is reportedly heading for a limited test in December, which could coincide with Google’s recently rumored launch of Gemini 2.0.
A report states Google is working to build a new AI agent that might make life feel like it’s on autopilot.
According to The Information (pay-walled), Google is working on a new “AI agent” and intends to implement it in Chrome (via 9to5Google). The publication cites persons “close to matter” in saying this AI’s name is Jarvis — Project Jarvis. This software can reportedly “respond” to a user’s commands by taking screenshots of their screen. While this sounds strange, the post states this is done to “interpret” what a user is doing before taking action.
Supposedly, the purpose of Jarvis is to satiate the itch some users may have with wanting more automation while online.
The post explains that, when taking “action,” Jarvis could click a button for you or fill in a text field with your required information. Those with inside knowledge have reportedly told the publication that Jarvis “operates relatively slowly” since the model needs time to process the (appropriate) next course of action.
Aside from filling in text fields, the report states Jarvis could be used for purchases and flight booking.
A concerning aspect of Jarvis as its (alleged) functions for automation is security. The publication highlighted this, stating how the AI can “safely handle” their information remains to be seen.
Though alarming, we’ll have to wait and see. The post claims Google is preparing an early preview of Jarvis that might be ready by December. This could arrive as a limited test for a small group of people.
This rumored timeframe of this “limited” test of Project Jarvis coincides with a report of Google’s next Gemini model. It was recently stated that Google is busy crafting Gemini 2.0, which is expected to launch in December. It’s unclear what advancements this 2.0 model will bring for users, especially after the company’s Gemini 1.5 Pro model. However, the report claims that Google may have made this decision due to its competitors like OpenAI.
The company is reportedly developing its next ChatGPT model dubbed “Orion.” Rumors state Orion may not launch this year, but a different version of ChatGPT could still debut in December.
During I/O 2024, Google clearly stated its focus on deleloping AI agents. As The Information highlighted, an AI agent is software that can act without a human operator/supervision. In May, Google said it was working on AI agents to help process information faster through video frame encoding. Moreover, it sees these agents as a way to help combine speech and video to create “a timeline of events and catch information relevant for recall.”