FEATURE
Google One AI Premium: AI Is Everywhere These Days
by Ryan Mulholand
AI seems to be in just about everything these days, doesn’t it? There’s AI in our search engines, AI in our chatbots, and, with time, AI will be in our doctors’ offices and restaurants. Since the launch
of ChatGPT in late 2022, AI has come to the forefront of everyone’s minds. This new and exciting tech has been used for everything—from creative media such as the “Studio Ghibli” craze on social media to helping advance novel therapies in cancer research and diagnostics. Everywhere we go, it seems as if everyone is developing their own AI models and putting them in their products and services [1, 2].
In February 2024, Google launched Google One AI Premium (one.google.com/about/google-ai-plans). This new service offers expanded access to Google Gemini’s Deep Research, as well as AI integration into many of Google’s already-existing services, such as Maps, YouTube, Docs, Sheets, Search, and Chrome. It also includes generative possibilities, such as video creation using Veo and NotebookLM. This is a premium service, and premium services have premium price tags. For $19.99 per month, you get access to the above-named services plus ones that haven’t been named.
This begs the question, “Is it really worth almost $20 per month?” Luckily for me and for other prospective users, Google One AI lets you try these services free for 1 month before starting a recurring subscription. It’s also advertising free access for students until final exams in 2026. With so many different services included, the opportunity to experiment was too good to ignore.
DEEP DIVE INTO RESEARCH
As a university research librarian, I figured I should start with Google Gemini’s Deep Research. Deep Research is a Google AI model in Gemini that delves deeper into whatever research topics you have. These deep dives include significantly more sources being cited as well as more advanced reasoning capabilities. I fed the model the following prompt about the cultural and intellectual influences of the French Revolution:
“I want a detailed report on the main intellectual competing forces that existed during the time of the French Revolution (1789–1799). Identify not only the main drivers of it, but also explain how the prevailing ideals ultimately helped contribute to the return of the monarchy, with the ascension of Napoleon I. Be sure to only use academic and scholarly sources. Popular sources like Quora are strictly prohibited.”
Gemini spent several minutes working through my request and returned a thorough report that was structured much like a college essay. The report detailed the origins of the ideas that would lead to the events of the French Revolution, as well as the big-name players and their respective factions. These people and groups were organized into easy-to-read charts that greatly enhanced the report’s readability, as well as explaining their cause-and-effect relationships. Once it was all said and done, the data used in the report provided by Gemini came from over 240 separate websites. These sources ranged from universities and archives to more popular publications such as Quora and Wikipedia.
Several popular publications appeared in the final write-up, even with the explicit prompting that such resources were forbidden. However, Gemini still returned and used resources from various academic publications. While the inclusion of popular sources was unfortunate, the use of multiple academic sources at least helps provide some measure of reliability. Overall, my time using Gemini’s more advanced features has been rather positive. If you’re someone who works in a data-heavy profession, then perhaps the features of Deep Research alone are enough to justify the price tag.
AI-ENHANCED GOOGLE PRODUCTS
Gemini’s more advanced features are just one part of the suite of offerings with a premium subscription. Next, I decided to explore what features were included that affect Google Docs, Sheets, and Slides. The benefits of using AI integration for the rest of Google’s products became noticeably less apparent from here on out, at least for me.
To really give Google Docs AI a test, I decided that the perfect exercise for it would be to help me rewrite my entire resume from scratch. My old one was getting crowded, and as anyone who has been through the rigors of job hunting knows, it’s not exactly the most fun or exciting thing to do. To make sure that I didn’t overwhelm the system, I input one section at a time into the prompt box and tasked Google with giving each section suggestions for improvement. Within 10 seconds, the built-in AI assistant returned suggestions for changes I could make and provided explanations for why I should make each one.
The total time that I spent rebuilding my resume took maybe an hour, and it now has a significantly improved structure, conciseness, and greater ease for future edits (which is especially helpful when needing to tailor a resume to a specific job posting). The feedback and suggestions I got from the built-in assistant were good, but I could have gotten similar results from the freely available versions of other large language models (LLMs). The free versions of other models or even Gemini could have done a similar job of helping me rebuild and revise my own resume. The use of a built-in assistant maybe saved me a little time by not having to alternate between tabs, but not much beyond that.
The tests with Google Sheets and Slides (Google’s version of Microsoft Excel and PowerPoint, respectively) left a similar impression to Docs. For Sheets, I tasked the assistant with helping me spot trends within a stock trading journal I keep for tracking profits and losses. The assistant was able to provide me with several formulae to help with my calculations, as well as provide a summary of my journal. Like my experience in Docs, I could have done all of this with a free model. As for Slides, I was able to use it to generate some basic images to act in support of a demo presentation I was throwing together, but I don’t see much use beyond that. Even prompting Gemini with possible use cases didn’t yield anything new or insightful.
Outside of my experiment with Docs, I really didn’t see much use for Google’s Microsoft suite equivalents. Maybe someone who uses both more regularly might, but for me, the features don’t justify the price. Sure, I could prompt Sheets to calculate sums, averages, and other calculations each time I need them. I could also simply look up the functions and commit them to memory with enough practice. There are still plenty of products that have AI built into them, so with this in mind, I turned my attention toward Google Maps.
GOOGLE MAPS
Maps presented Google’s AI with a chance to redeem the underwhelming performance of some of the other apps. Within my Maps app, the AI assistant provided a brief summary of one of my favorite restaurants—Angela’s Miami Grill. The summary of the restaurant was brief and to the point, as well as accurate as to the kinds of food that Angela’s serves. Scrolling a little down the features list of the restaurant, Google provided a prompt box that allowed me to ask questions about the location and its services.
Since Angela’s is a place that my colleagues and I love to gather once a month to discuss issues and fraternize, I decided to ask the assistant how much a beer was to gauge its accuracy and reliability. The assistant told me that there wasn’t any available information on the prices of beer (normally about $5 per pint) and that if I wanted to know, I would have to call and ask. While the pricing of alcohol wasn’t readily available, I was able to get an estimate for a plate of food ($10–$20 per entree, according to the AI).
I was able to repeat these results with several other restaurants in the area, with the only real issue being getting a price quote on alcohol. Even then, the assistant was able to answer the most basic questions I had about an eatery. While some restaurants will more readily have information such as pricing available, using an AI assistant is still a better system than using several dollar signs to denote how expensive a place is/can be.
GOOGLE AI OVERVIEWS
My last experiments were on Google Search. Most of the people who have used this recently have probably seen brief summaries being generated when they input their search criteria. These summaries are a part of Google’s AI Overviews, an algorithm that is designed to try and gather information from multiple sources and then present it in an easy-to-read manner. These summaries tend to be generated for people, places, and events. In my own line of work, I will use these summaries when meeting one-on-one with students seeking research help to quickly get contextual knowledge about a topic.
While these overviews are certainly no substitute for deeper research, the benefit of it is in its conciseness. The brief (two to three) paragraphs written in Overview can be what sparks greater interest in a subject. I’ve used these overviews for everything from historical figures like Nietzsche to events like the Congress of Berlin to scientific concepts such as aposematic coloration (bright colors mean “danger”). This feature is probably what I have used most out of all of the different AI integrated features for Google.
There are other apps that integrate Google AI, but I feel as though the ones we’ve covered are more than enough. So, is the $20 per month price tag worth it? For me, as things presently stand, no. As I have already noted, I could already do pretty much everything that the built-in assistant can do, but with a free model. With different services and tools, I can freely generate images (Ideogram), summarize videos (NoteGPT, Glasp), and brainstorm solutions to problems (really, any LLM), to give a few examples.
VALUE FROM GOOGLE GEMINI
The biggest value for me comes in the form of expanded access to different forms of Google Gemini. The Deep Research model has been a great personal research tool for exploring different subjects of interest to me. The exploration of literally hundreds of different websites (550 is the most I’ve seen in one research report) at the very least shows a wide reach for information. The addition of showing its real-time thought process also helps me better understand how the system works.
Beyond the expanded access to Gemini, the rest of the features don’t exactly appeal to me. Perhaps someone who heavily uses Google products and services might see better use of a built-in AI assistant, but for me, this would simply save a little time switching between tabs for freely available apps.
Outside of even those considerations, the sheer volume of services included in this premium subscription make Google One AI unlike any other AI service I’ve seen to date. Within this one package, you can get more details about a restaurant, edit your favorite photos, schedule appointments and events on your calendar, and so much more. This service is also still fairly new and still in its infancy.
With all of that in mind, I’d be more than willing to re-evaluate these services after about a year. The beginnings of something better are there; the service just needs time to develop somewhat to grow into its potential. |
