Confessions of a “Recovering” Data Broker: Responsible Innovation in the Age of Big Data, Big Brother, and the Coming Skynet Terminators

  • 654 views
Uploaded on

It's been said that the human brain is comprised of 300 million pattern matchers fed with data from our five primary senses and memories. In this age of distributed computing and cheap storage in the …

It's been said that the human brain is comprised of 300 million pattern matchers fed with data from our five primary senses and memories. In this age of distributed computing and cheap storage in the cloud, "thinking" without a biological brain is possible for the first time in history. The sensory input into this new, extracorporeal brain is big data.

Global data supply chains carry exabytes of government, corporate, and social data powering breakthrough uses in medicine, transportation, communications, and energy. However, equally fantastic is the specter of abuses by powerful players to exploit private information, subtly discriminate, or mistakenly prosecute the innocent.

This talk discusses the current state of these data supply chains, where they are headed, and the societal implications for privacy, security, and liberty. And it calls technologists, business leaders, and humanists -- i.e., geeks, suits, and wonks -- to together resolve the tension between cultural values and fast-paced technology.

More in: Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
No Downloads

Views

Total Views
654
On Slideshare
0
From Embeds
0
Number of Embeds
3

Actions

Shares
Downloads
11
Comments
0
Likes
2

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide
  • NSA sphereBut it's for a good reason -- keep us safe. Security and privacy play the roles of the pig and the chicken at breakfast. Like the pig, security requires commitment, Privacy, like the chicken, is merely involved.Privacy loses to security by default … unless we are vigilant.For security, the peril is often felt immediatelyWith privacy, the peril is often felt much later. Why we discount privacy perils so much – the present cost of a distant future disaster is zero.Privacy and Civil LIberties Oversight Board. (PCLOB)Data, information, knowledge, wisdomNext: Not so sure this is wise …Ref“An individual who breaks a law that conscience tells him is unjust, and who willingly accepts the penalty of imprisonment in order to arouse the conscience of the community over its injustice, is in reality expressing the highest respect for the law” ― Martin Luther King Jr.

Transcript

  • 1. Confessions  of  a  "Recovering"  Data  Broker     Responsible  Innova.on  in  the  Age  of  Big  Data  and  Big  Brother   Jim  Adler   Vice  President,  Products   Metanau.x     jimadler@metanau.x.com   @jim_adler   hDp://jimadler.me     Markkula  Center  for  Applied  Ethics   Feb  25  2014    
  • 2. Plea “Can’t  we  all  just  get  along?”   −  Rodney  King   Geeks   High-­‐Tech   Mercenary   Social   Entrepreneur   Responsible   Innovator   Suits   Tradi?onal   Capitalist   Wonks  
  • 3. Lesson Eclectic  generalists  drive  innovation.   Richard  Feynman   Quantum  Physics   Steve  Jobs   ‘nuff  said   Stephen  Hawking   Cosmology   Norio  Ohga   Sony  President   74  min  CD   Temple  Grandin   Animal  Handling  
  • 4. Confession I  am  not  an  attorney.   Intelligence   Geek   Obsession   Dweeb   Nerd   Dork   Social   Inep.tude  
  • 5. Confession You  can  often  do  more  good  from  the   inside  than  the  outside.   •  Founded  in  2003   •  20B  public  records   •  30M  visitors  per  month   •  50M+  reports  sold  
  • 6. Confession The  “public”  data  supply  chain  of  you   Payments   Civil   Suits   Criminal   Records   Commercial Risk   Resumes   Government Public   Posts   Names   Addresses   Blogs   Search   Phone   Numbers   Collection Self-Reported Big Data Engines Marke.ng   Directory   Background   Use
  • 7. Confession We  don’t  know  you  all  that  well.   Billions  of  Records   Millions  of  People   Philip   Collins   375  People   Jim  Adler   213  Records   37  People   Carol  Brooks   9800  Records   1250  People   Randolph   Hutchins   5  People   Gwen   Fleming   2  People   213  records  linked   to  the  correct  37  Jim  Adlers     Jim  Adler   Houston,  TX   Age  70   Jim  Adler   McKinney,  TX   Age  57   Jim  Adler   Has.ngs,  NE   Age  32   Jim  Adler   Canaan,  NH   Age  59   Jim  Adler   Redmond,  WA   Age  50   Jim  Adler   Denver,  CO   Age  48  
  • 8. Confession BANKING   SERVICES   Lots  of  uses  for  your  data  …  some  regulated.   CALLER  ID  OF  HARASSING   PHONE  CALLS   ONLINE  SHOPPERS   VERIFYING  ONLINE  SELLERS   LEARNING  ABOUT     ADOPTED  KIDS  SEEKING  THEIR   BIOLOGICAL  PARENTS   A  BUSINESS   SOCIAL  NETWORKERS  LOOKING  TO   EXPAND  THEIR  FRIENDS  LIST   NON-­‐PROFIT  ORGANIZATIONS   LOOKING  FOR  SUPPORTERS   CHECKING  OUT  A   PROSPECTIVE  TENANT   THOSE  IN  LEGALLY  ENTANGLED   LOOKING  FOR  COURT  RECORDS   RESEARCHING  A   PROSPECTIVE  EMPLOYEE   ALUMNI  GROUPS   ARRANGING  REUNIONS   NETWORKERS  SEEKING   BUSINESS  OPPORTUNITIES   GENEALOGISTS    CULTIVATING   THEIR  FAMILY  TREE   ANYONE  RETRIEVING   COURT  RECORDS   SHARING   FIND  OWNER  OF  DOG’S   RELATIVE  FOR  TRANSPLANT   RESEARCH   SALES  PROFESSIONALS  LOOKING   FOR  NEW  PROSPECTS   CHECKING  OUT  A   PROSPECTIVE  DATE   FIANCÉS  AND  THEIR  CURIOUS   FAMILY  MEMBERS   ENFORCEMENT   SOCIAL  WORKERS  WHO  NEED  TO  KNOW   MORE  ABOUT  THEIR  CLIENTS   LAWYERS  NEEDING  QUICK  ACCESS   TO  COURT  RECORDS   BUSINESSES  THAT  NEED  TO  UPDATE  CONTACT   INFORMATION  ON  CUSTOMERS   ANYONE  CURIOUS  ABOUT  WHO'S   EMAILING  OR  CALLING  THEM   AIRLINES  TRYING  TO   RETURN  LOST  LUGGAGE   PROFESSIONALS  LEARNING  ABOUT   COLLEAGUES  AT  CONFERENCES   SINGLES  CURIOUS  ABOUT  THE   PEOPLE  THEY  MEET   INVESTIGATIVE  JOURNALISTS   RUNNING  DOWN  LEADS   RECONNECTING  OUT-­‐OF-­‐TOUCH   FAMILY  MEMBERS   LAW   PARENTS    ENSURING  WHO   THEIR  KIDS  SAFETY   FINDING  PEOPLE  THAT  HAVE  THE   SAME  ILLNESS  AS  YOU   ANYONE  WHO  NEED  ADDRESS   HISTORIES  FOR  PASSPORTS   CHECKING  OUT  A  PROSPECTIVE  SOCIAL   NETWORK  CONNECTION   FINDING  LONG-­‐LOST  FRIENDS,  MILITARY   BUDDIES,  ROOMMATES,  OR  CLASSMATES   REGULATED  
  • 9. Confession Opt-­‐out  doesn’t  always  mean  deletion.   Jane Hampton Jane Hampton Jane Hampton 06/23/1998 123 Main Peoria, IL 123 Main Peoria, IL jane@facebook.com (309)555-8931
  • 10. Lesson Listen  to  your  toughest  critics.  
  • 11. Lesson Can  I  have  a  little  narcissism  with  my   voyeurism?   •  What  does  my  background   check  say?   •  Privacy  controls   –  Suppress  single  address  or   phone  number   •  Comment  on  your  own   public  profile  
  • 12. Regulation New  Data  Broker  Bill  Introduced  This  Month   •  “Data  Broker  Accountability  and  Transparency  Act”   •  Prohibits  “data  brokers”  from  using  decep?ve  means  to  collect   informa.on  about  consumers.   •  Transparency  to  consumers  about  informa.on  about  them.   •  Consumers  can  correct  the  data.   •  Opt-­‐out  of  having  their  data  collected.   •  FTC  enforcement  
  • 13. Lesson When  towns  were  small,  personal   anonymity  was  low  …   “The  only  thing   worse  than  being   talked  about,  is   not  being  talked   about.”     −  Oscar  Wilde  
  • 14. Lesson Urban  populations  grew  along  with   personal  anonymity…   “Good  Fences   Make  Good   Neighbors”     −  Robert  Frost  
  • 15. Confession …  we’re  suffering  from  Privacy  Vertigo.   120   “Rockwell”  Era   “Good  Fences”  Era   Privacy Expectations ! 100   “Privacy   Ver.go”   Era   80      Onli     nsity ne  De 60   40   20   Urban  Density   0   1850   1890   1930   1970   2010  
  • 16. In  privacy  contexts,  Power  matters.   Privacy  Rights  !   Lesson Peer  to  Peer   Corpora.on  &   Customer/Employee   Government  &   Ci.zen   Power  Disparity  !     Your  God  &  You  
  • 17. Lesson How  to  unpack  Privacy?  Think  PPP.   PERILS  
  • 18. Mapping  Places-­‐Players-­‐Perils  Cases   Private Curtilage Governments Employers/Landlords/Insurers Public M O R E   P L A Y E R   P O W E R   G A P   Lesson Parents Peers M O R E   P R I V A T E   P L A C E S  
  • 19. Lesson Places-­‐Players-­‐Perils  Cases   M O R E   P L A Y E R   P O W E R   G A P   US  deports  Bri?sh   tourists  over  Tweets   NSA  internet  ci?zen   surveillance   Georgia  teacher   FBI  GPS  criminal   fired  aXer  pos?ng   Google  privacy   surveillance   vaca?on  pics   policy  unifica?on   "Girls  Around  Me"   pulled  from  market   Ethically Challenging Target  finds  out   News  of  the  World   teen  pregnant   Health  orgs  use   phone  hacking   before  parents   Twi[er  to  track   illness   Actress  sues  IMDB   over  revealing  her   age   FB  user  sets  fire  to   home  aXer  de-­‐ friending   GM  OnStar  tracks   users   M O R E   P R I V A T E   P L A C E S   Woman  caught   naked  by  Google   Street  View   Rutgers  student   commits  suicide   aXer  spied  by   webcam  
  • 20. M O R E   P L A Y E R   P O W E R   G A P   Confession Big  brother  is  watching  (duh).   NSA  internet   “We’re  being  asked  to  trust   US  deports  Bri?sh   ci?zen   without  being  able  to  verify.”   tourists  over  Tweets   surveillance   −  Alex  Howard  (big  data  journalist)   GPS  criminal   Georgia  teacher   FBI   fired  aXer  pos?ng   Google  privacy   vaca?on  pics   policy  unifica?on   surveillance   Target  finds  out   News  of  the  World   teen  pregnant   Health  orgs  use   phone  hacking   before  parents   Twi[er  to  track   illness   Pres.  Obama  calls  for  more   transparency  in  FISA  court  and   Woman  caught   Actress  sues  IMDB   surveillance  laws   naked  by  Google   "Girls  Around  Me"   pulled  from  market   over  revealing  her   age   FB  user  sets  fire  to   home  aXer  de-­‐ friending   GM  OnStar  tracks   users   Street  View   Rutgers  student   commits  suicide   aXer  spied  by   webcam   NSA  chief  announces  plan  to   replace  1,000  sysadmins  with   machines   M O R E   P R I V A T E   P L A C E S  
  • 21. Lesson Technology  grows  exponentially.     Wisdom  grows  linearly.   •  Gov’t  doesn’t  trust   people  (at  least   sysadmins)  but  does   trust  machines   •  LiDle  Transparency     •  Wisdom  is  hard  to   come  by     •  Sen.ent  (?)  brain  in   the  cloud  in  <  20   years   Wisdom   Knowledge   Informa.on   Data  
  • 22. Prediction A  head  in  the  clouds  <  20  years   $100,000   Human  Brain   •  20,000  TFlops   •  2,500  Terabytes   More  than  $325M  per  year   $27,100   $13,500   Cost  per  Month  (000s)   $10,000   $6,800   $3,400   Chris  Westbury,  University  of  Alberta     Less  than  $700K  per  year   à  4  M  AWS  m1.large  nodes   $1,700   $1,000   $850   $420   $210   $100   $100   $53   $26   $13   $10   $7   $3   $2   $1   $1   2012   2014   2016   2018   2020   2022   2024   2026   2028   2030   2032   2034   2036   2038   2040   2042   Year  
  • 23. Lesson Big  data  inferences  are  not  thoughtcrimes.   “…  the  essen.al  crime  that   contained  all  others  in  itself.   Thoughtcrime,  they  called  it.”       –  George  Orwell   “Watch  your  thoughts,  they  become  words.   Watch  your  words,  they  become  ac.ons.   Watch  your  ac.ons,  they  become  habits.   Watch  your  habits,  they  become  your  character.   Watch  your  character,  it  becomes  your  des.ny.”      –  Lao  Tzu  
  • 24. Confession Target  knows  you’re  pregnant  and  when   you’re  due.  So,  what’s  so  perilous?  
  • 25. Confession “To  Serve  Man”  is  a  cookbook.   “If  you’re  not   paying  for  the   product,  you  are   the  product.”     −  Claire  Wolfe   (paraphrased)  
  • 26. Sometimes  you’re  in  a  public  place  when   you  think  you’re  in  a  private  place.   Confession “Gaydar” A 2009 MIT study found it was possible to predict men’s sexual orientation by analyzing the gender and sexuality of their social network contacts – even if the rest of the information on their profile was set to private.
  • 27. Confession John  Foreman’s  Excellent  Disney   Adventure  
  • 28. Felon  Classiier   Sampling   Linking   250  M   Defendants   Cleaning   Objec.ve   If  someone  has  minor  offenses   on  their  criminal  record,     do  they  also  have  felonies?   15K  Labels   15K  Predictors   Feature   Extrac.on   Learner   Model   Bloomberg  ar.cle:  hDp://bloom.bg/1eMtnug    
  • 29. How  does  the  Felon  Classiier  work?   Gender                                   Eye  Color   Ta[oos   Male  (+0.1)   Blue   2  +   Female   Brown  (+1.2)   <  2  (+0.1)   Green   Criminal  Offenses   Score   Over   Threshold   of  3.5?   Likely     Felon?       Traffic  only  (-­‐0.5)   4  or  fewer   misdemeanors   8  or  fewer   misdemeanors   YES   0.9 NO   NO   Hazel       Male   Female  (-­‐0.5)   Blue   2  +  (+1.3)   Traffic  only   Brown   <  2   4  or  fewer   misdemeanors  (+1.9)   Green   8  or  fewer   misdemeanors   YES   4.4 NO   YES   Hazel  (+1.7)   Bloomberg  ar.cle:  hDp://bloom.bg/1eMtnug     Blog  widget:  hDp://jimadler.me    
  • 30. Confession Classiiers  depend  on  policy  as  much  as   technology.   False  Negative  Rate   A N A R C H Y   100.0%   80.0%   60.0%   40.0%   Threshold:  1.1   FP  Rate:    1%     FN  Rate:  40%     Threshold:  0.66   FP  Rate:    5%     FN  Rate:  22%     Threshold:  -­‐1.82   FP  Rate:    19%     FN  Rate:  0%     20.0%   0.0%   0.0%   5.0%   10.0%   False  Positive  Rate   T Y R A N N Y   15.0%   20.0%  
  • 31. Ruling NYC  Stop  &  Frisk  Found  Unconstitutional   90% of Criminals are Minorities Minorities 50% Criminals “The  city  …  believes  that   blacks  and  Hispanics   should  be  stopped  at  the   same  rate  as  their   propor.on  of  the  local   criminal  suspect   popula.on.”   −  US  District  Judge  Shira  Scheindlin     All NYC Residents
  • 32. Lesson “Half  the  money  I  spend  on  advertising  is  wasted;   the  trouble  is  I  don't  know  which  half.”   Bayes’  Rule   PMinority is a Criminal = Minorities 50% 10% of Criminals are Not Minorities Criminals 5% 90% of Criminals are Minorities PCriminal is a Minority PCriminal PMinority PMinority is a Criminal = 90% 5% = 9% 50% PMinority is NOT a Criminal = 100 − PMinority is a Criminal = 91% All NYC Residents If  it’s  not  ok  to  stop  99%  of  the  general  popula.on  for  nothing,   why  is  it  ok  to  stop  91%  of  minori.es  for  nothing?  
  • 33. Lesson When  might  Stop  &  Frisk  be  OK?   Bayes’  Rule   PMinority is a Criminal = Asians 10% Criminals 5% PCriminal is a Minority PCriminal PMinority PMinority is a Criminal = 90% 5% = 45% 10% PMinority is NOT a Criminal = 100 − PMinority is a Criminal = 55% All NYC Residents If  it’s  not  ok  to  stop  99%  of  the  general  popula.on  for  nothing,   is  it  ok  to  stop  55%  of  minori.es  for  nothing?  
  • 34. Lesson What  about  Newark’s  Stop  &  Frisk?   hDp://www.ny.mes.com/2014/02/25/nyregion/newark-­‐ stop-­‐and-­‐frisk-­‐data-­‐is-­‐analyzed.html     Bayes’  Rule   Minorities 50% Criminals 20% PMinority is a Criminal = PCriminal is a Minority PCriminal PMinority PMinority is a Criminal = 90% 20% = 36% 50% PMinority is NOT a Criminal = 100 − PMinority is a Criminal = 64% All Newark Residents If  it’s  not  ok  to  stop  96%  of  the  general  popula.on  for  nothing,   is  it  ok  to  stop  64%  of  minori.es  for  nothing?  
  • 35. Lesson Hilary  Mason’s  Maxim   Math  +  Code  =  Awesome   Quants Making  a  killing  on  Wall  Street  but  s.ll  can’t  impress  the  chicks   Weakonomics.com  
  • 36. Lesson Corollary  to  Mason’s  Maxim   Values  *  (Math  +  Code)  =  Awesome  
  • 37. Prediction “The  beatings  will  continue  until  morale   improves.”   Overton  Window   Unthinkable Acceptable Radical Facebook Newsfeed Ad Targeting Facebook Beacon Pre-crime Skynet Popular Sensible Policy
  • 38. Lesson Living  within  a  Filter  Bubble   (with  apologies  to  Eli  Pariser)   …  but  then  we   reshape  our  tools  …   “We  shape     our  tools  …     …  and  thereawer   our  tools  shape  us.”   −  Marshall  McLuhan  
  • 39. Lesson “No  one  here  gets  out  alive.”   −  Jim  Morrison   Listen,   Learn   Geeks   Suits   Wonks   Scru.nize,   Incen.vize   Adapt,   Invent  
  • 40. Questions?     Jim  Adler     www.metanau.x.com   jimadler@metanau.x.com   @jim_adler