The recent Google I/O event left a lasting impression, with its abundance of product reveals, features, and announcements. Amidst the multitude of presentations, one recurring term stood out: “AI.” Interestingly, someone took the initiative to tally its utterance count—149 times, to be exact—without relying on AI. The prominence of AI was particularly evident in the demonstration of remarkable tools capable of drafting emails with just a brief prompt and assisting users in Google Workspace. These advancements left us yearning to experience each AI-powered feature firsthand.
Nevertheless, as Google endeavors to catch up with the proliferation of generative tools by introducing its own AI offerings, some existing apps and services seem to have been left behind. However, they could greatly benefit from the implementation of artificial intelligence. While we appreciate Google’s apps and services, there are occasions when they fall short of meeting our expectations, leaving us somewhat frustrated. We believe that before delving into more sophisticated endeavors, Google should employ AI to address certain day-to-day challenges.
1. Gboard’s Typing Woes
In a world where AI can compose comprehensive articles from a simple prompt, it remains perplexing that accurate prediction of typed text can still prove challenging. Gboard, Android’s default keyboard, is widely regarded as one of the best available. However, it occasionally exhibits unpredictable behavior and becomes unreliable.
For instance, as a fan of swipe typing due to its speed, I’ve noticed Gboard frequently failing to recognize the correct word. Additionally, it has recently developed a peculiar habit of randomly capitalizing words in the middle of sentences. This forces me to backtrack and rectify the issue, nullifying the anticipated speed advantage. Multilingual users, who often have multiple Gboard languages enabled, face an even greater challenge.
Furthermore, the text suggestion bar occasionally recommends words that defy grammatical sense within the context of the sentence being typed. Although Google has made continuous efforts to refine its AI prediction algorithms over the years to account for human errors while typing swiftly or swiping, the persisting issues indicate the need for further improvement.
2. Gmail’s Writing Assistance
The introduction of the “Help me write” feature for Gmail during Google I/O was an exhilarating moment, second only to the introduction of the dark theme for Bard. However, when we take a step back, we realize that our Gmail inboxes are inundated with thousands of unread and unnecessary junk emails. While Gmail does offer filters to help sift through the clutter—for example, displaying only promotional emails from specific brands received after a certain date—many users remain unaware of these tools. Furthermore, the syntax required to configure them proves too complex for those lacking technical expertise.
Artificial intelligence has the potential to assist billions of Gmail users by automating sorting based on email content and recommending the deletion of outdated items, as demonstrated in the Files by Google app. Google could allow users to create their own automation recipes while also providing preset options for everyone to neatly organize their email inboxes.
3. Restoring Intelligence to Smart Speakers
Voice assistant-powered smart speakers continue to be a fixture in our lives, albeit with a slight decline in attention from Google (and Amazon) compared to a couple of years ago. Similar to Gboard, the Google Assistant experience varies significantly across users and devices, leaving these so-called smart speakers feeling less intelligent lately.
There have been numerous instances where I asked my Google Home speaker to perform a task, and it responded by playing a random song from YouTube Music. It would mistakenly set an alarm for the evening instead of the morning, provide a rephrased version of my query while still failing to offer any relevant information, or even repeat the same answer it had previously given. The list of frustrations continues to grow with each interaction.
Google Assistant on smart speakers could greatly benefit from the impressive AI capabilities showcased in newer Google products. These smart speakers should make our lives easier instead of causing frustration when attempting to execute simple tasks.
4. Enhancing Google Photos’ Search Functionality
Google Photos is undeniably the most intelligent photo management app, with a significant portion of its appeal stemming from its robust search tool. However, the search feature has remained largely unchanged since its initial introduction several years ago. While Google recently introduced complex searches that allow users to locate specific themes, such as people at the beach or colorful sunsets, after all this time, Photos should be capable of even more.
A major drawback of Photos’ search function is its reliance on a limited set of standard keywords. Using substitute words or phrases often yields different or incomplete results, causing users to miss out on relevant images. For example, while it’s possible to search for birds in general, Photos struggles to identify specific bird species. The same issue arises when searching for a particular make and model of a car, as Photos frequently confuses and intermixes the results.
The manner in which queries are typed can also significantly affect search results. Given the complexity of languages and their variations across different regions, even for widely spoken languages like English, it is crucial for Photos to understand context in order to provide accurate results. Considering that Google Search has been accomplishing this feat for years, implementing a similar approach within Photos should be feasible.
5. Curbing the Nuisance of Spam Calls and Texts
Spam calls and texts have become a modern-day nuisance. While Google offers a few useful tools to combat them, such as Call Screening and the spam alert integrated into the company’s dialer app, both solutions have their limitations. Call Screening is often unavailable on non-Pixel phones and in many countries, while the reliability of spam alerts remains questionable.
There are instances when the Google Phone app mistakenly marks important calls as spam, making it difficult to rely on its spam detection. Consequently, users are left at the mercy of third-party tools like Truecaller, which, although more dependable, are known for their invasive privacy practices in managing spam calls.
Spam has also infiltrated Google Messages, with brands employing RCS to flood inboxes with intrusive advertisements. While there are workarounds, user-side solutions struggle to keep pace with the ever-evolving tactics employed by spammers. By implementing automated and more capable spam management measures, akin to our earlier suggestion for Gmail, Google could alleviate the inconvenience associated with spam calls and texts for many Android users.
The Rapid Advancement of Google’s AI and its Product Offerings
A proficient artificial intelligence system should continuously learn, adapting to user behavior and interactions to enhance the overall user experience. Predictive AI has long been an integral aspect of various Google services, including Assistant and Gboard. However, this new era of generative AI has sparked our excitement, as it generates content seemingly from scratch, displaying an uncanny realism.
While these two AI applications differ significantly, the self-awareness inherent in Google’s generative AI initiatives can greatly benefit the traditional predictive AI we commonly encounter. For instance, Gmail could create custom automation recipes based on user behavior, and Google Photos could identify people based on specific outfit details. We eagerly anticipate witnessing such improvements in our favorite Google services, as we await transformative AI applications that surpass our present imagination.