Push AI Progress, Here's A Row Of Google Innovations In April

JAKARTA Google has always launched the latest innovation in Artificial Intelligence (AI). This development shows that Google's investment in machine learning and AI research is successful.

April is one of the busiest months according to Google as they introduce a lot of AI innovations. They introduce a simple text converter into a short cinematic video to a multimodal search into AI Mode.

Of the many AI innovations Google has revealed, some of them also include Gemini's upgrades. Here are a number of AI innovations that Google announced last April.

Gemini 2.5 Pro and Agent Deep Research

Google finally switched from model 2.0 Flash to 2.5 Pro last month. Although still experimental, Gemini's early adoption received enthusiasm from the public, especially app developers.

The reason is, the 2.5 Pro version can be used to work on coding. Google has also launched an initial version of Gemini 2.5 Flash for developers at Gemini API via Google AI Studio and Vertex AI.

After launching the 2.5 Pro version which is still experimental, Google presents a personal research assistant called Deep Research. Supported by Gemini 2.5 Pro, this AI model can support research thanks to detailed source searches.

Camera Support and Image Editoring at Gemini

The next upgrade is camera support and screen sharing capabilities on Gemini Live. With this feature, users can ask or discuss it in real-time while activating the camera. Users can enable the camera to the desired object.

If you use the screen sharing mode while Gemini Live is active, users will show what they are looking at on the screen. Gemini will read the screen and users can ask questions from what is being viewed.

The following ability that is no less important is AI editing by uploading images directly to Gemini. This feature was originally only available to AI Studio users, but was eventually extended to Gemini in April.

Latest Experimental Cybersecurity Model

The next innovation Google announced last April was Sec-Gemini v1, an experimental AI model designed to improve cybersecurity. This feature will reduce the burden facing cybercriminals.

Sec-Gemini v1 can double the strength of professionals in cyberspace. Judging from its name, Google is likely to launch the next generation of this cybersecurity AI model.