NEXT: Conversations with Alexa
If I want to know last night’s baseball scores, I can type my query into Google. Up will come a number of search results, from ESPN.com to CBS Sports Network, to the Major League Baseball site, to the sites for the local teams’ newspapers. Getting to the top of that search is, of course, big business.
Now that I have an Alexa or some other smart speaker in my home, I can now “speak to the room” and ask “Who won last night’s game?” Alexa will return a result immediately, “Washington beat Los Angeles 7-3 in 10 innings.”
Alexa does not have that answer in memory, of course. The algorithm searches the web, as I had previously done, locates or otherwise puts the information together and presents it to me. As long as I learn the baseball result, I guess it does not matter to me that the score comes from ESPN or the MLB site. I just want the answer.
Well, as you can imagine, it does indeed matter to the companies that want your eyeballs fixed on their sites. Like the Oracle at Delphi, we are increasingly coming to expect “instant, perfect answers” to our queries, what some are calling “one-shot answers.” Think of the computer on Star Trek: The Next Generation. Crew members would speak aloud some question, for which the computer—in that pleasant woman’s voice—would immediately respond with the correct answer. And indeed, it would appear that many of us are treating our smart speakers exactly as if they were the Star Trek computer: asking our smart speakers to answer a question is the principal way most of us use these devices.
It is called “the conversational web.”
Many companies are now formulating their conversational web strategies. When I ask Alexa to tell me the score of the game, she/it is not telling me the source of her information. Indeed, her source might come from a complex taxonomy—Google’s is called the Knowledge Graph—that organizes many sources of information. If all I want is an answer to my query, where the information comes from may be of less interest to me. The next competitive space for many companies will involve making certain that the answer the smart speaker provides us will be coming from their information.
For Alexa to be so “conversational,” complex knowledge ontologies have been constructed, first by Google and other large tech firms. But it may soon be the case that libraries will become an important provider of such knowledge ontologies. After all, librarians have long been in the business of organizing knowledge. A consortium of libraries might draw from their vast collections to become the most sophisticated provider of conversational web services.
If the web is becoming conversational, it is so far a kind of one-sided conversation. It is possible that smart speakers, powered by artificial intelligence, will one day allow us to ask more sophisticated questions. Alexa would develop a broader and more complex range of responses. In such a scenario, will we start to have real conversations? If conversation is defined as dialogue between two or more people where news and ideas are exchanged, might we start to have proper conversations with Alexa? Instead of a one-sided “do this” or “give me the answer to this” kind of conversation, might we instead exchange ideas and ask Alexa, “Do you think Andrew Yang’s proposal for a Universal Basic Income is feasible and politically viable?” or “How might we combat climate change?” or “Is behavior innate or learned?” or “Will the U.S. and China competition end in warfare?” Alexa might become a conversation partner as we puzzle through complex, and perhaps even unanswerable, questions.
There are important implications for schooling. The logic of standardized testing insists that students be able to recall answers to questions upon demand. If answers can be so readily generated by artificial intelligence, then perhaps the mark of an educated person in the future will include the ability to formulate and ask good questions.
We might see a new form of Socratic dialogue emerge between human and artificially intelligent interlocutors.