The Software Society

How digital technology is changing our culture and economy

Can Artificial Intelligence create a new non-technical job category?

As the Wall Street Journal put it in its May 3 headline, “Job Growth Gathers Strength,” but with the caution that there were “numerous worrisome signs, among them another exodus of workers from the labor force and persistently weak wages.” Technologies such as speech recognition and natural language processing can reduce the number of jobs that formerly required human skills (e.g., medical report transcription and call center agents). What new category of well-paying jobs might these language technologies create to compensate for those lost? There are, of course, highly technical jobs in language technology itself that require years of education or experience. But is there a new job category that doesn’t require years of training?

One potential such category will evolve as individuals get accustomed to natural-language personal assistant apps, such as those offered by Apple, Google, and Microsoft. The underlying technology can support specialized “conversational apps” that can be a new form of entertainment, marketing, and/or information delivery, but will involve intensive human involvement to create the original content and to support on a continuing basis. The key human skill required is creativity or knowledge about a specific content area, rather than computer skills.

Today, entertainment and information applications are largely at two extremes of user interaction: (1) intensively interactive (e.g., games); or minimally interactive (e.g., enter a search term to get an answer or watch a video or TV show). A conversational app is an intermediate option. Such an app would involve an intermediate level of interaction with the user, with all categories of response to the user an option (images, text, and/or audio). Categories of conversational apps include:

  • Entertainment: A video or audio story/game that requires repeated interaction with the user to progress, with the story/game changing depending on what the user says. For example, one app might be a detective story where the user asks questions to get clues and solve a mystery. Another might be a children’s story where the story changes based on preferences stated by the child.
  • Information and education: An interactive application where the user can inquire about a subject within the declared range of the application, e.g., investing for retirement accounts. To go beyond search, a true dialog to refine an inquiry should be possible, with retention of the context of previous dialog (allowing inquiries such as “How do I do that?”). Options can include guidance through a series of steps (e.g., executing a recipe), with the user indicating when the next step should be provided, and optionally asking for more detailed instructions.
  • Marketing: Marketing apps can be a combination of entertainment and information. They will typically be branded and represent a particular firm or product. Options can include descriptions of products, frequently asked questions, options to buy, technical support, customer service, or all of these. The customer may be encouraged to use the app by an entertaining introduction, e.g., differing funny responses in reaction to a customer utterance within the context of the company product line (e.g., tasteful jokes about the competition).

A fundamental characteristic of conversational apps is that they branch, that is, they are not a linear, fixed experience such as a book or movie. And they don’t have the fixed rules that define a game. As such, they require much more content than more conventional media. Also, many of the genres listed admit of continual addition of content to add new material or to better respond to a user inquiry that only gets an evasive answer. A cooking assistant can evolve into an encyclopedia on cooking, for example.

By their nature, most conversational apps will require human creativity. Much human time will be required to create a conversational application and support it. Many conversational applications will require a team.

Many jobs and entrepreneurial opportunities will be created by conversational apps. An attraction of such jobs is that creativity or expertise in specific topics (e.g., hobbies) is all that is required. One doesn’t have to take a course in computer science to create such an app if the appropriate tools are available—content is king. Hopefully, tools and web sites will appear to support those who feel they can create conversational apps. It will only take a few successful conversational apps to start a deluge, as the rapid growth in conventional mobile apps has proven.

4 thoughts on “Can Artificial Intelligence create a new non-technical job category?

  • Deborah Dahl says:

    This is a very thought-provoking post. Thinking about the kinds of tools that would be needed to support this kind of ecosystem, some of them are already starting to become available in ways that non-experts might be comfortable working with — especially speech recognition and natural language understanding. But two more categories of tools will be essential. One is tools for creating dialogs — the back and forth interaction between the system and the user that makes an application happen. Making dialog creation easy is going to require building the basic principles of cooperative dialogs into tools so that developers can focus on content, not on the mechanics of conducting a dialog. The dialog part needs to just work. The second category of tools we need is tools that make it easy to bridge the interface between the application and existing content such as the vast amount of structured information already on the Web. This information is often available from API’s like those found on the Programmable Web website (over 11,000 API’s currently listed!), but accessing these API’s can be complicated for non-experts. Semantic Web resources will also become a rich source of structured content. Just as writing got to be a lot easier after writers stopped having to make their own paper and ink, creating intelligent, content-rich, applications will be much easier when the basic tools and infrastructure are already in place and creative people can just build on them.

    • Deborah,
      Good points. On the need for dialog creation tools, I wonder if open chatbot platforms like the recently launched BOT Libre! (http://www.botlibre.com) might be leveraged to create a sort of Content Management System to support content creators for virtual assistants. For example, it seems that creating a platform where authors could create interactive stories for virtual assistant to “tell” wouldn’t be that complicated. The interaction would be somewhat contrived, but at least the story would be engaging.
      In terms of the tools for discovering existing content on the web, I wonder if the markup defined under schema.org might be a place to start? But I’m not an expert on semantic web. There may be better options.
      It’s interesting that IBM Watson now has a “debater” capability, which seems to mean that it can quickly identify relevant content on the Internet to formulate arguments for and against a position (apparently without using markup of any kind).
      Amy

  • I like the concept of interactive conversational apps and I agree that this area has lots of potential. The 3 conversational app categories that you list are good ones. I’ve also been thinking along the lines that it would be great to have a platform that would allow story tellers to easily create interactive story content that could be consumed by any ‘standard’ virtual assistant, perhaps via an API or something. But how would content providers profit from their content? Can we envision a near-term future where a marketplace for virtual assistant content exists that is accessible by virtual assistant vendors? There would need to be a way for the virtual assistant vendor to access the content seamlessly, a way for the user to pay for the content, and a way for the content provider to get reimbursed. I don’t know how the virtual assistant marketplace will evolve, but perhaps this is a possible future scenario.

  • […] Meisel noted that one trend creates a major new category of creative jobs. “Since conversation can engage an individual more tightly than simply listening or reading, we will see a new category of digital and web applications that converse with us, and adapt their responses to our responses. This is, for example, likely to eventually be a major trend in advertising, what the Wall Street Journal called ‘chatvertising.’ And interactive entertainment and information sources that require much more material than a linear book or movie will probably require a team of writers, providing more jobs for creative people.” (See Can Artificial Intelligence create a new non-technical job category?) […]

Leave a Reply

Your email address will not be published. Required fields are marked *