Jump to content

  • Log in with Facebook Log in with Twitter Log In with Google Log In with Steam Sign In
  • Create Account
Photo

Amazon Alexa is now crowdsourcing answers from random users


  • Please log in to reply
No replies to this topic

#1
OFFLINE   sincity

sincity

    Advanced Member

  • Members
  • PipPipPipPip
  • 2496 posts
63
Getting Better

Starting today, Amazon is trialling a new crowdsourcing program for its Alexa smart assistant - one that will look to users like you to fill in the gaps of its knowledge. (No pressure.)

The Alexa Answers program has begun inviting select customers to take part in the beta, which allows you to browse through unanswered questions on various topics - science, technology, trivia, and the like - and offer your two cents. 

Those answers are then offered up to other users who can vote the suggestion up or down in order to rate its accuracy or helpfulness. There's no claim to fame for upvoted answers, sadly, but Alexa will let you know if its used "an Amazon customer" as a source.

There are naturally filters for keeping out profane language; there will always be those determined to muck around, though we imagine Alexa will be able to prevent 'unhelpful' users from submitting ever more answers.

Help a brother out

For a smart assistant whose main use is in offering information to those who need it, Alexa sure can get things wrong.

Alexa uses a variety of sources for gathering information from across the web - via Bing searches, online dictionaries, or crowdsourced sites like Wikipedia - with the margins for error that entails.

The key thing to remember is that the internet is littered with mistakes and misinformation, and that inevitably trickles down to the answers users find on Google, Wikipedia, social media, and naturally AI like Alexa or Siri. Amazon already relies on human-sourced reviews for rating and advertising its products, so the Alexa Answers program is in some ways just a natural extension of that philosophy.

Whether adding another human element to the mix will make things better, or worse, will depend on how motivated users are to help.

Via The Verge






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users