{"id":313027,"date":"2017-06-16T02:44:47","date_gmt":"2017-06-16T01:44:47","guid":{"rendered":"http:\/\/rinf.com\/alt-news\/newswire\/facebooks-a-i-chatbots-begin-communicating-after-inventing-new-language\/"},"modified":"2017-06-16T02:44:47","modified_gmt":"2017-06-16T01:44:47","slug":"facebooks-a-i-chatbots-begin-communicating-after-inventing-new-language","status":"publish","type":"post","link":"http:\/\/rinf.com\/alt-news\/newswire\/facebooks-a-i-chatbots-begin-communicating-after-inventing-new-language\/","title":{"rendered":"Facebook\u2019s A.I. \u2018Chatbots\u2019 Begin Communicating After Inventing New Language"},"content":{"rendered":"<div>\n<p>      <!-- START CONTENT --><\/p>\n<p>Researchers noticed a \u201cdivergence from human language\u201d during chatbot interactions<\/p>\n<p><strong>Mikael Thalen<\/strong><br \/>Prison Planet.com<br \/>June 15, 2017<\/p>\n<p>Chatbots created at the Facebook Artificial Intelligence Research lab developed their own language without being directed by\u00a0researchers.<\/p>\n<p>According to a <a href=\"https:\/\/s3.amazonaws.com\/end-to-end-negotiator\/end-to-end-negotiator.pdf\" target=\"_blank\" rel=\"noopener noreferrer\">report<\/a> released Wednesday, the\u00a0discovery was made during a project that <a href=\"https:\/\/www.yahoo.com\/news\/facebook-gives-bots-ability-negotiate-compromise-175629309.html\">gave bots<\/a>\u00a0the ability to negotiate and make compromises.<\/p>\n<p>As the bots\u2019 development progressed, researched say they suddenly noticed a\u00a0\u201cdivergence from human language\u201d during interactions, forcing them to alter their model.<\/p>\n<p>\u201cIn other words, the model that allowed two bots to have a conversation\u2014and use machine learning to constantly iterate strategies for that conversation along the way\u2014led to those bots communicating in their own non-human language,\u201d <a href=\"https:\/\/www.theatlantic.com\/technology\/archive\/2017\/06\/artificial-intelligence-develops-its-own-non-human-language\/530436\/\">notes<\/a> The Atlantic\u2019s Adrienne LaFrance.<\/p>\n<p>A similar scenario was seen last year when the \u201cGoogle Translate\u201d service <a href=\"https:\/\/www.newscientist.com\/article\/2114748-google-translate-ai-invents-its-own-language-to-translate-with\/\">invented<\/a> its own language to translate with.<\/p>\n<p>\u201cThe online translation tool recently started using a\u00a0neural network\u00a0to translate between some of its most popular languages \u2013 and the system is now so clever it can do this for language pairs on which it has not been explicitly trained,\u201d New Scientist\u2019s Sam Wong <a href=\"https:\/\/www.newscientist.com\/article\/2114748-google-translate-ai-invents-its-own-language-to-translate-with\/\">wrote<\/a>e. \u201cTo do this, it seems to have created its own artificial language.\u201d<!-- [donotprint] --><\/p>\n<ul class=\"ppani\">\n<li>A d v e r t i s e m e n t<\/li>\n<li\/><\/ul>\n<p><!-- [\/donotprint] --><\/p>\n<p>Most bots prior to Wednesday\u2019s announcement were largely limited in their ability to carry out complex conversations, researchers say. Now, Facebook\u2019s chatbots are not only capable of\u00a0estimating and negotiating over an item\u2019s\u00a0\u201cvalue,\u201d\u00a0but can even use deception in order to broker a deal.<\/p>\n<p>Researchers said the\u00a0bots \u201cinitially feigned interest in a valueless item, only to later \u2018compromise\u2019 by conceding it \u2014 an effective negotiating tactic that people use regularly\u201d \u2013 a method developed by the bots on their own.<\/p>\n<p>    &#8230;<\/p><\/div>\n<p><a href=\"http:\/\/www.prisonplanet.com\/facebooks-a-i-chatbots-begin-communicating-after-inventing-new-language.html\">Read more<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Researchers noticed a \u201cdivergence from human language\u201d during chatbot interactions Mikael ThalenPrison Planet.comJune 15, 2017 Chatbots created at the Facebook Artificial Intelligence Research lab developed their own language without being directed by\u00a0researchers. According to a report released Wednesday, the\u00a0discovery was made during a project that gave bots\u00a0the ability to negotiate and make compromises. As the [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[519],"tags":[],"class_list":{"0":"post-313027","1":"post","2":"type-post","3":"status-publish","4":"format-standard","6":"category-newswire"},"_links":{"self":[{"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/posts\/313027","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/comments?post=313027"}],"version-history":[{"count":0,"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/posts\/313027\/revisions"}],"wp:attachment":[{"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/media?parent=313027"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/categories?post=313027"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/rinf.com\/alt-news\/wp-json\/wp\/v2\/tags?post=313027"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}