Meta testing Hatch AI assistant as AI race with OpenAI, Anthropic heats up: Report

Wait 5 sec.

Meta is reportedly developing Hatch, an AI agent designed to automate tasks and compete with premium offerings from OpenAI and Anthropic. (File photo: Jason Henry/The New York Times)Meta is reportedly preparing to expand its artificial intelligence (AI) ambitions with a new AI agent called Hatch, as the company looks to compete more aggressively with rivals such as OpenAI and Anthropic.The tech giant is also considering a tiered pricing structure for Hatch, and the proposed pricing would place Hatch alongside premium plans such as ChatGPT Pro and Anthropic’s Claude Max, both of which are currently priced at $200 a month.The reported move comes as Meta intensifies its focus on AI, with CEO Mark Zuckerberg positioning the company as a stronger competitor in the rapidly evolving AI space. Although Meta has lagged behind some rivals in advanced large language models, the company appears to be betting heavily on AI-powered agents as a key growth area.Also Read | Do AI bots now outnumber humans online? Here’s what Cloudflare’s Matthew Prince saysUnlike conventional chatbots, AI agents are designed to complete tasks on behalf of users. These systems can be instructed to handle activities such as writing code, drafting presentations, scheduling meetings or sending emails with minimal manual input.Reports suggest Hatch is being developed as a consumer-facing AI agent that may function similarly to OpenClaw, a tool that gained attention earlier this year for its ability to automate digital tasks locally on a user’s device.Test versions of Hatch reportedly allow users to describe tasks in plain language, after which the AI can create software tools or execute actions independently. One example cited in reports involved the system building a fitness tracker based on a user request.In addition, Hatch may include a dashboard or feed where users can view and manage tools created by the AI. Meta is also said to be exploring modular “skills” that could allow the agent to automate workflows, connect with third-party services and perform specialised tasks such as creating travel itineraries.Story continues below this adAlso Read | OpenAI rolls out Lockdown Mode to protect against prompt injection attacksFurthermore, reports indicate that Meta could introduce a “Hatch Plus” tier that offers significantly higher daily usage limits than the free version. Similar to competing AI platforms, usage allowances may reset each billing cycle and may not carry forward.The pricing strategy suggests Meta is positioning Hatch as a premium offering rather than competing on affordability, signalling a more aggressive push into the AI assistant market.Unlike standalone AI platforms that require separate sign-ups, Hatch could reportedly be integrated into Meta-owned services, including Instagram, potentially giving it access to billions of existing users.Meta is also believed to be testing Hatch using Anthropic’s Claude Sonnet models for now, although the final product is expected to run on Meta’s in-house Muse Spark AI model.Story continues below this adThe launch timeline, however, remains subject to change as Meta continues testing the product. The proposed AI agent is part of Zuckerberg’s broader push to turn Meta’s massive investments in AI into sustainable revenue streams. In recent months, the company has expanded premium Meta AI subscriptions and reorganised internal teams around AI-focused development.(This article has been curated by Shivani P Menon, who is an intern with The Indian Express)