a.i.

Marketing in the Digital Age: News Round-Up (week ending Jun 8, 2018)

Sipra-Favicon-NoStroke.jpg

In this week's Marketing in the Digital Age, several subjects of A.I. have come up and Facebook has a breach of data, again.  With A.I. rapidly progressing and coming into the forefront of media, what type of regulation needs to be in place to keep humanity protected? 

Facebook Breaches Data of its Users

Summary: New York Times reported this week that Facebook gave 60 device makers access to users' data and that of their friends without express permission.  Facebook denies this.  However, they have confirmed that a data-sharing partnership with Chinese firms including Huawei (a company flagged by US intelligence as a security threat).  Apparently the information transferred was stored only on a user's device, not Chinese company servers.  

Opinion: A lot of what has been happening in the social media space has gone unchecked, largely because it's a new arena and the implications from it haven't been planned for or  contemplated.  I imagine that will quickly change with the massive breaches Facebook has undergone the past several months.  

Google Won't Use A.I. for Evil

Summary:  Google CEO Sundar Pichai announced yesterday an emphasis of Alphabet's philosophy of "don't be evil" when he confirmed that Google won't use its A.I. tools for weapons or surveillance.- with some caveats.  It will work with the government and military on cybersecurity and surveillance but nothing that violates "internationally accepted norms".   

Opinion: What is one woman's/man's evil is not necessarily the other's.  A.I. is coming, but as you see from the Facebook breaches, a level of regulation also needs to be set here to ensure humanity's safety.  Using a vague avoidance of what violates "internationally accepted norms" could likely vary on more subtle points from one person to the next (or one country to the next country).

A.I. Training in the Dark Corners of Reddit

Summary: Researchers at MIT trained an artificial intelligence psychopath machine, Norman (as in Norman Bates), using Reddit.  The purpose was to show that the data used to train machine learning algorithms influences its behaviors.  “Norman suffered from extended exposure to the darkest corners of Reddit,” the researchers state, “and represents a case study on the dangers of artificial intelligence gone wrong when biased data is used in machine learning algorithms.” Example of what Norman responded with: 

inkblot.png

 

Opinion: See opinion above.  Don't let it get to be a problem before potential solutions and regulations are explored! 

@tinycarebot vs. @infinite_scream

Summary: Two Twitter bots have been stuck on a seemingly endless loop of "conversation" for hours on Twitter.  @infinite_screeam will respond to any tweet with a poignant "AAAAAAAAAAHHHHHHH" while @tinycarebot sends out uplifting, emoji-filled messages. So every uplifting tweet is met with virtual screams.  

Opinion:  Hah, just a stupid and fun way to go into the weekend before A.I. takes over writing these blogs for me.