Google is bringing image generation directly into AI Overviews in Search, allowing people to create a new visual from a text prompt without first moving to a separate image-generation product. The company is also introducing a redesigned Google Images home built around a personalised, continuously updated gallery.

The two changes were announced on 14 July 2026 as Google marked 25 years since the launch of Google Images. Although the anniversary provides the backdrop, the more significant product development is the continuing convergence of search, visual discovery and generative AI inside Google's core consumer services.

Nano Banana moves into AI Overviews

Google says image generation in AI Overviews will use its latest Nano Banana model. A user will be able to provide a text description and receive a newly generated image within the Search experience, addressing requests where a suitable image does not already exist on the web.

The feature is due to begin rolling out over the coming weeks in English. It will be available in regions that already support image creation in AI Mode, meaning the announcement is not a simultaneous worldwide release. Google did not provide a country-by-country timetable, detailed usage limits or separate pricing information in the announcement.

Putting image creation into AI Overviews reduces the distance between asking a question and producing an asset. That could make generative imagery more accessible to people who would not seek out a dedicated creative tool. It may also encourage users to treat Search as a place to create information and media, rather than only locate material that already exists.

A more personalised Google Images home

The second update is a new browseable home for Google Images. Google describes it as a dynamic gallery assembled from images across the web, updated in real time and tailored to a user's interests. Saved ideas and collections will appear as tabs above the main gallery so users can return to themes they have previously explored.

This redesigned experience has a narrower initial release. Google says it will roll out over the coming weeks on desktop in the United States and in English. A Google Account sign-in is required. Those conditions are important: the feature is not yet a general replacement for the existing Google Images experience in other markets, languages or on mobile devices.

The personalised gallery also raises practical questions that the announcement does not answer in detail. Google did not explain how people can inspect or adjust the interest signals used for recommendations, how frequently personalisation can be reset, or how publishers may experience changes in referral traffic. Existing Google Account and Search controls may apply, but organisations should verify the final product documentation as the feature becomes available.

Visual search becomes a creation interface

Google positioned the new features as the next stage in a long shift away from text-only search. The company's visual search history includes reverse image search, Google Lens, combined text-and-image queries, Circle to Search and AI Mode features that analyse an entire scene. Google says Circle to Search is now available on more than 580 million Android devices, illustrating the scale at which visual input has already entered everyday search behaviour.

The addition of image generation changes the role of the interface. Traditional image search ranks and links to existing material. A generative result creates a new output based on a user's instructions. That distinction matters for provenance, expectations and evaluation: a generated image should not be treated as documentary evidence, even when it appears beside conventional search results.

Businesses may find the capability useful for early concepts, presentation illustrations, mood boards and rapid visual exploration. It does not remove the need to review generated material for accuracy, brand suitability, intellectual-property risk and unintended bias. Google did not announce enterprise governance controls, indemnity terms or guarantees for generated outputs in this consumer Search update.

Availability and unanswered questions

  • AI Overviews image generation: rolling out over the coming weeks in English where image creation in AI Mode is already supported.
  • New Google Images home: rolling out over the coming weeks on desktop in the United States, in English, for signed-in users.
  • Model: Google says the generation feature uses its latest Nano Banana image model.
  • Pricing and limits: no separate price, quota or detailed eligibility schedule was included in the announcement.

As the rollout proceeds, users should look for clearer labelling between web-sourced and generated images, controls for personalisation, information about data retention and model-improvement use, and guidance on commercial use. These details will shape whether the feature is mainly a convenient consumer tool or also becomes useful in governed workplace processes.

AI Provider Index assessment

This is a meaningful distribution move rather than a new foundation-model launch. Google is placing generative image creation inside a high-frequency discovery workflow, which may expose the capability to a broader audience than a standalone application could reach. At the same time, the staged rollout and missing operational details make it too early to judge consistency, control or suitability for professional production.

For the wider AI market, the update reinforces a pattern in which large platform providers embed generation into existing products instead of asking users to adopt an entirely new destination. Competing search, assistant and creative platforms will increasingly be compared not only on model quality, but also on where generation appears, how clearly it is labelled and what controls surround it.