Google: Ending the Search for Artificial Intelligence (AI)

Part of the inhumanity of the computer is that, once it is competently programmed and working smoothly, it is completely honest.

- Isaac Asimov

In the past few weeks, Google have announced three products that aim to help us be more productive: Priority Inbox, Scribe, and Instant. Not only do they cut down on the time it takes to process information and do repeat tasks, but they aim to start predicting the thought processes operating in our minds. Combined, these three applications are slowly bringing an end to the search for artificial intelligence.

Priority Inbox is a tool for GMail that sorts out your messages into various layers of importance. When you first turn on the feature, it will take a guess as to what you deem important. It was fairly accurate in separating my messages out, and continually gets better as I train it further by marking messages as important or not important. This feature could be duplicated by using filters and labels, but a key difference is that the Priority Inbox will start sorting new mail as it arrives. With filters, you generally create them after the mail is in your inbox already. Priority Inbox allows us to focus on one inbox of important messages instead of sifting through messages and labels, letting the unimportant messages stay hidden until we have time to work with them.

Scribe takes the idea found in Priority Inbox of guessing as to what we think is important by looking at our previous work or popular works on the Internet to create sentences. You enter one word into Scribe and it begins a rapid search to find the next logical word or words in the sentence. Not only does it find the word that fits correctly, but it also tries to create a voice in the writing – your voice. Scribe is a beta feature right now and does work incredibly well at times, but then library ad infinitum offers the uncanny turing test of Google Scribe to show the other side of it. For example, “In the beginning, God created the world and the world of the living room and dining room with a view to share videos with friends and family to enjoy a good meal.”

Instant is a new search feature that tries to present us with the search results we are most likely looking for. It only works with a Google account, and the results are based upon a combination of what we have searched for in the past, as well as, the most popular searches on the Internet. It presents the results to you live, refreshing with every new key stroke, and the results include images and maps. Instant is incredibly fast and provides excellent results in guessing at my search entries.

The combination of these three services aids us in our ability to analyze the data flow heading towards us, and to take action upon it. Over at the Singularity Hub, Aaron Saenz questions whether Priority Inbox is a new form of AI assistant, because it helps manage the flow of inbound information. “We’re are going to be flooded with incoming information as billions of people around the world jump online in the next few decades… AI personal assistants will be able to manage this information with simple filtering, allowing these data streams to enhance your experience rather than overwhelm it.”

While “enhancing the [user's] experience” of interacting with data is important, it is also important that these new AI personal assistants behave in unobstrusive ways. The Google tools will only work if they disappear into the background and become a subconcious with our data interactions. The best example of this can be found in auto-complete functions, from typing in email addresses, passwords, to using a program like TextExpander to populate frequently typed text. These functions have changed how we work with those text fields, and we use them without even thinking about it. They have become part of our nature in entering in data.

Priority Inbox has already generated a lot of discussion about how necessary it is and how much easier managing one’s email inbox has come (Leo Laporte, an infamous Apple Mail supporter, even migrated over to GMail because of this feature). That is impressive for a tool that has only been released to the public the past two weeks. A similar move will happen with Instant as it gets released into the mobile search realm and people become more used to the pre-generated search results.

As our level of comfort grows in trusting these subconcious processes, a new level of intelligence will begin to appear. Our minds will adapt to focus on creation and interactivity; machines will learn how to do more nuanced tasks with our data. The more we hand things over to computers, the more data will be collected on what exactly we mean by an “approrpiate response,” or how to respond to an “urgent request.” Eventually, machines will become to learn that Justin Bieber is an important figure in the social realms of teenagers, and by analyzing what they write about him, start to converse with them directly. When I start a sentence in Scribe with “Justin Bieber,” I get, “Justin Bieber is a Canadian singer and songwriter who has been in the business of making music for about a year.” I can envision within five years that the computer can write a decent article just by entering in a subject and the length.

The discussion data that Google is pulling together is only one piece of a large puzzle they are slowly putting together. People often wonder why Google is branching out into so many different areas, or criticize them for including so many features in GMail, but it’s becoming more clear what the end goal is becoming.

Think about the services they currently have or will be unveiling in the next year:

  • GMail/Talk/Reader/Blogger/News – how we form sentences, what items are important enough to read and interact with, how we converse in short form and long, what are the best ways to present an argument when writing.
  • Search/Instant – what we think is important to know, what is popular, when do we want the information, how often do we seek it out.
  • Images/YouTube/Music – what is popular to look at, to watch, what are effective ways of presenting data in a visual/audio/both format, what do we enjoy listening to.
  • Maps/Places/Latitude – where do we like to go, how do we get there, what makes a business popular, when do we travel.
  • Android, GoogleTV/Music, and GoogleMe allows them to collect data around the clock as we are going to be tuned into Google in some means during the majority of our time.

How they bundle all this data into a usable form to develop an AI machine is another question, but they could be working in the background on solving that problem. An article published three and a half years ago suggested that Google was developing an AI. Now, we are at that tipping point where the data is extensive enough to create a high quality AI machine.

The future is, truly, now.