Virtual Assistant Responses Are Carefully Crafted by People
The top names in tech now each have digital assistants, with Amazon, Apple, Facebook, Google and Microsoft each offering their own flavour of virtual advisors that listen and respond to voice commands and keystrokes. Apple introduced Siri in 2011, and Microsoft’s Cortana, Google’s Now and Amazon’s Alexa were rolled out several years later. Facebook M is still in beta stages.
It feels odd to talk to a piece of software, but, in reality, these utilities are more human than we realise. Responses are carefully crafted by specialists, with the team behind Cortana, for example, consisting of novelists, screenwriters, playwrights and essayists. The authors review each response and decide whether or not it needs revising. Artificial intelligence allows the assistants to learn from these changes, and the data is fed back to the database to automate more-accurate responses.
Writers decide whether answers are meaningful for people from different cultures, backgrounds and age groups, but tend to remain gender-neutral in their responses and avoid any political leaning. Users often ask inappropriate questions, to which writers create polite and meaningful answers.
The goal of those creating virtual assistants is to raise them to be trusted and respectful. They’re developed to exhibit human-like, multidimensional intelligence, as usage is based on their perception. It takes respect to build a relationship.
The complex engines behind virtual assistants are a combination of basic automation and “deep learning” technology — artificial intelligence that can infer the user’s intent to make suggestions and plan actions. It’s prediction science. None of the current platforms are perfect, but responses have gradually become more helpful, accurate and even friendlier. As virtual assistants grow up, they’re maturing into complex interfaces. It takes a village to raise a child, but a team of specialists to raise an intelligent personal assistant.
Twitter
Facebook
LinkedIn
Email