SIAgentNews.com
← All news

Google makes its Spark AI agent free for users

Google is rebranding its Spark agent as Tasks and offering it to free users, providing a low-cost entry point into autonomous AI task management.

· First reported by ZDNet

100Mweekly token limit for Meta's Muse

The short version

  • Google renamed the Spark AI agent to Tasks and made it available to free users in the US, Canada, Korea, Japan, Australia, and India.
  • Tasks runs on the Gemini 4 Hydrogen model and can perform actions like booking appointments, grocery shopping, and analyzing bills.
  • The agent includes a feature for personal finance, such as identifying recurring expenses and finding deals.
  • Users can start tasks via email or text, and the agent can call businesses for them in most US states.
  • Meta's Muse agent is also free but has a weekly limit of 100M tokens.

Why it matters

This move lowers the barrier to entry for personal AI agents, which can autonomously handle complex real-world tasks. It allows casual users to experiment with agentic technology without committing to expensive monthly subscriptions.

The shift to Tasks

Google is officially rebranding its Spark AI agent to Tasks, making the tool available to users who do not pay for a premium subscription. The transition began gradually in mid-October, and users can now find the new agent by toggling from the standard Gemini Chat settings. This move follows the retirement of Gems, which were previously used to tailor Gemini for specific jobs.

The rebranding comes as the tech industry moves toward personal agents that can perform actions autonomously. While paid versions of these agents exist, Google is positioning Tasks as a more accessible option for the general public. By offering it for free with specific limits, Google aims to capture a wider audience of users who want convenience without high costs.

The new landing page for Tasks highlights specific use cases like personal finance and cost-cutting. It encourages users to let the AI identify recurring expenses they no longer need or set up alerts for desired deals. This shift suggests Google wants to move beyond simple chat and into active life management for its users.

How the agent functions

Tasks operates on Google's newest Gemini 4 Hydrogen model to handle a variety of complex requests. It is designed to perform proactive jobs such as booking appointments, grocery shopping, and even analyzing a user's heating bill. These capabilities allow the agent to interact with the digital and physical world on behalf of the user.

Users in the United States can initiate these tasks through various channels, including sending an email or a text message to Gemini. In many cases, the agent can even place phone calls to businesses to complete a request. However, there are geographic restrictions on the calling feature, which is unavailable in Indiana, Louisiana, Minnesota, Montana, and Nebraska.

To manage the data involved in these interactions, Google provides tools for user control. The company states that users can adjust task settings, disconnect third-party apps, or delete their chat history at any time. Additionally, a new Memory summary feature allows users to view and edit the personal information the chatbot has stored about them.

The competitive landscape

Google is entering a crowded market where other tech giants are also developing personal AI agents. For example, OpenAI offers Dots agents as part of a $100 per month ChatGPT Pro tier, while Anthropic provides Claude Duet starting at $20 per month. These paid options currently set a high price point for high-end agentic capabilities.

Meta has also entered the space with its Muse agent, which is free to use but limited to 100M tokens per week. Because Muse was previously described as having significant privacy concerns, some users may prefer Google's Tasks as a less risky alternative. Both Muse and Tasks are currently the most affordable personal agents available for those who are hesitant to pay for full access.

The availability of these free options suggests that the technology is still maturing. Since the paid agents have not yet fully impressed many users, the free versions provide a way for the public to test the tech. This allows users to see if the convenience of an autonomous agent outweighs the potential risks of giving an AI permission to handle personal tasks.

Privacy and security risks

Because AI agents can autonomously handle tasks like placing orders or reviewing private files, they present unique security risks. Users can set spending caps on these actions to mitigate some of the financial risks associated with autonomous spending. However, the ability for an AI to proactively manage a user's life means that privacy remains a primary concern for many potential adopters.

Google has attempted to address these concerns by providing transparency through the Memory summary feature. This tool is intended to give users a clearer picture of what the AI knows about them. By allowing users to edit this information, Google hopes to build trust with people who are wary of how their data is used to power these agents.

The move toward agentic products is driven by a society that increasingly relies on technology for productivity and convenience. As these tools become more integrated into daily life, the balance between ease of use and data security will be a major factor for users. For now, the free tiers of Tasks and Muse serve as a testing ground for how these systems will eventually operate on a larger scale.

What comes next for agents

As Google expands the reach of Tasks, the focus will likely shift toward how well these agents can handle increasingly complex real-world interactions. The current free model allows Google to gather data on how people use agents for common tasks like shopping and scheduling. This feedback will likely inform future updates to the Gemini 4 Hydrogen model and the Tasks interface.

Competition from Apple, OpenAI, and Anthropic will continue to push the boundaries of what personal agents can do. While Google is currently focusing on personal finance and basic logistics, other companies may offer more specialized agents. The market is moving toward a model where users can choose between high-cost, high-capability agents and free, limited-use versions.

For the average user, the next few months will be a period of experimentation. With Tasks now available in major markets like the US, Canada, and Japan, more people will be able to see if the agent can actually save them time. The success of these tools will depend on whether they can provide genuine utility while maintaining a level of privacy that users find acceptable.

Try it

Check your Gemini Chat settings to see if you can toggle the Tasks agent on and try asking it to identify recurring expenses in your emails.

Watch

How Google handles privacy as more users grant Tasks permission to access third-party apps.

Whether OpenAI or Anthropic will introduce free-tier agents to compete with Google's Tasks.

If the calling feature for Tasks expands to the five US states where it is currently restricted.

Sources and links

First reported by ZDNet. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news