“The era when AI does the job for you is not far away,” said an expert.
From video generation to report summarization and data analysis, the potential applications of AI are expanding rapidly.
Upon its release, I personally used it to see how deeply generative AI can integrate into actual work tasks. Google’s introduction of the latest model, Gemini Advanced 2.5 Pro, last March marked a significant stride in AI technology, offering a level of execution distinctly different from existing AI services.
The document summarization capability was impressive. The range of the context window stood out the most, as it could summarize entire reports or theses while maintaining their structure, even with complex documents containing code, tables, and lengthy descriptions. The support for one million tokens was palpable in practical use, and future expansion to two million tokens raised expectations even further.
The video generation model Veo 2, applied from April, proved functional albeit brief. An 8-second video at 16:9 aspect ratio and 720p resolution was generated from a single line of text. Even with simple prompts like “city night view,” the results were natural. Although the video quality exceeded expectations, Google plans to limit the number of monthly generations due to the function’s substantial computational demands.
The camera and screen sharing functions are activated on Android. When pointing the camera at an actual object, Gemini immediately recognized and provided information on it. Responses were prompted by book covers, objects, and landscapes, demonstrating the possibility of functions connecting information search to real actions even outside the office.
The Deep Research function is based on Gemini Advanced’s experimental model, provided to subscribers of the 2.5 Pro version. Upon entering a desired topic, users receive a report summarizing the core content, which can also be listened to in audio form. This feature facilitated preparation for presentations and research tasks by enabling repetitive learning even while on the go. It felt like receiving immediately usable information rather than just conducting a simple search.
Gems are customizable AI tailored to the purposes set by the user. Recently, all Gemini users have access to this feature. For instance, by creating a “health trainer” Gem, it can suggest daily exercises and meal plans based on dietary habits and exercise records. It is suitable for tasks requiring repetition or personalized recommendations.
Data analysis capabilities have also been expanded. Google Sheets, CSV, and Excel files can be uploaded, and data cleaning, analysis, and visualization can be handled within a single interface. The analysis results can be exported back to Google Sheets, enhancing real-world applicability.
Gemini Advanced is available through the Google One AI Premium plan, which includes 2TB of cloud storage and integrated workspace functionalities for 29,000 won per month. Gemini can be executed directly from Google Docs, Sheets, and Slides, allowing for a seamless workflow.
Gemini is more than just an AI for exchanging conversations. It is a generative AI platform encompassing content creation, data analysis, video generation, and real-time responses, now reaching a level where specific roles can be assigned to it.