Google Philippines’ Gemini era is in full swing as it conducted its own AI workshop for content and media creators Thursday, August 29. Complete with a lunch menu partly generated by artificial intelligence, the event sought to showcase the full capabilities of Gemini AI, Google’s successor to Bard.

Gemini AI is a versatile tool for answering questions, generating images, and summarizing and visualizing data. But, it’s not magic. In early 2024, Google’s AI overview caught flak for erroneous hallucinations promoting glue in pizza, which it has since been attempting to rectify.

As a creative day-to-day assistant, however, Google is keen to show that Gemini AI’s use case potential is vast and diverse, given the right input.

Here are some tips to improve the quality of your prompts and get the best results in Gemini AI and Gemini Advanced, which comes with the Google One AI subscription.

1) Let artificial intelligence take on real roles

Specificity is key to getting the most out of AI feedback. One way to contextualize prompts is by letting Gemini take on personas. For instance, you can brainstorm ad campaign ideas by starting with “Act as a social media manager ” or “expert product manager marketing skin care for Gen-Z audiences.” Be as rich with details as you wish, including brand messages, desired outcomes, target audience, and formatting directions. Avoid jargon and acronyms.

2) Create quick comparisons exportable to Google Sheets

Gemini AI is integrated with capabilities in search and data visualizations exportable to Google Sheets. This combination allows you to create custom comparisons, be it for gadgets, hotels, or other products.

With Gemini Advanced, it’s possible to upload documents, like tables and reports, to unearth personalized insights and visualizations.

3) Summarize documents

Gemini Advanced users get access to its next-generation model, which can process up to 1 million tokens, enabling a longer context window and analysis of up to 1,500 pages of a PDF file. Users can receive summarized insights on trends, such as year-by-year report assessments.

4) Generate images for slides

Instead of using stock photos, AI can generate unique scenes for presentations. Gemini is integrated into Slides, as well as other Google offerings like Docs and Gmail. In Slides, it can also easily remove image backgrounds.

5) Translate meetings with live captions

Starting January 22, 2025, translated captions in Google Meet will be exclusive to Gemini Enterprise, AI Meetings and Messaging, and Gemini Education premium. Gemini is available in 40+ languages, like Arabic, Portuguese, and Korean. However, according to Google, Gemini Advanced does not yet support any Filipino languages.

6) Get inspiration from existing photos

The upload tool enables users to gather insights from existing images.

For example, a business owner can ask the chatbot for advice on maximizing purchases by rearranging its in-store layout. A designer can upload snapshots of an office space and let Gemini Advanced break down the design aspects that make them pleasing to the eye. Breaking down what makes an image “work” can aid in crafting mood boards and design pegs.

In our experience, however, this feature is still pretty faulty especially in the free version of Gemini AI. 

7) Pin, edit, and iterate

Once a prompt is submitted in the input box and Gemini yields an unsatisfying answer, you can edit and resubmit your prompt. You can also iterate using natural dialogue to add details and contexts to the conversation. Pinning the particular chat in the sidebar allows easier navigation for returning to important conversations.

8) Double-check responses

At the end of every output, a Google button appears at the bottom. When clicked, it provides links supporting factual statements, as well as highlights sentences that may have had different or no results on the internet.

“If you have doubts about the credibility of its answer, you can also check sources in Gemini,” says Marl Pizzaro, a food and lifestyle vlogger. For him, AI is invaluable for getting quick explain-like-I’m-ten explanations and researching amidst a vast pool of online resources.

“I have a stern and bland tone on work emails and I am using AI to add more warmth and make things more casual and communicative,” said Coach Abby, a content creator focusing on technology and education, “Writer’s block is a deterrent to consistency and having Gemini to brainstorm with and provide another perspective is another thing that helps me make content.”

“On the creator space, I appreciate the safety features with the watermarking via Synth ID. AI should not take the place of creatives and artists, so being able to determine which is AI-generated and which one is human-made is essential.”

Synth ID is Google’s digital watermark. It’s essentially a hidden signature embedded within AI-generated content, such as images, text, and audio. It is imperceptible to the human eye, meaning it won’t affect the quality or appearance of the content. However, it can be detected to combat misinformation and deep fakes.

Gemini AI can be accessed through this link. Its premium version is available with Google One AI for P1,100.– Rappler.com