U4ab Wang Tao

Wang Tao Ms. Park ENGL 106 4/5/2014 Artificial intelligence A f!t!re "ro#le$ for %!
$anit&'
Wit% t%e fast (e)elo"$ent of t%e tec%nolog&* artificial intelligence+A,- #eco$e $ore an( $ore "o.erf!l. /ince , a$ st!(&ing 0o$"!ter Engineering. , .ant to fin( o!t t%e f!t!re t%e A,. W%en "eo"le %ear( a#o!t A,* first co$e to $in( is a ro#ot. As .e (esire $ore on t%e f!nctions t%at a ro#ot can (o* .e "!t $ore an( $ore $oralit& into t%e$. T%en it .ill lea( to t%e 1!estion t%at relate to t%e $ac%ine et%ics. As t%e $ore $oral t%e ro#ot #eco$e* .ill t%e& so$e(a& #eco$e !ncontrolla#le an( acti)ate( #& t%ere o.n .ill' ,f .e caref!ll& eno!g% to (esign t%e$* it s%o!l( not #e a "ro#le$ t%at $ig%t res!lt in a (isastro!s iss!e. What is the problem? ,t see$s like ro#ot re#ellion onl& %a""ens in t%e science fictions. 2o.e)er* after $& e3tensi)e researc%. T%e "ro#le$ are real e3ist in t%e .orl(. Accor(ing to an article .ritten #& 0ollin Allen4 Mac%ine et%ics isn5t $erel& science fiction. it5s a to"ic t%at re1!ires serio!s consi(eration* gi)en t%e ra"i( e$ergence of increasingl& co$"le3 a!tono$o!s soft.are agents an( ro#ots. Mac%ine et%ics is an e$erging fiel( t%at seeks to i$"le$ent $oral (ecision $aking fac!lties in co$"!ters an( ro#ots. We alrea(& %a)e se$ia!tono$o!s ro#ots an( soft.are agents t%at )iolate et%ical stan(ar(s as a $atter of co!rse. ,n t%e case of A, an( ro#otics* fearf!l scenarios range fro$ t%e
f!t!re takeo)er of %!$anit& #& a s!"erior for$ of A, to t%e %a)oc create( #& en(lessl& re"ro(!cing nano #ots+A#stractW%ic% $eans t%at t%e "otential factor of t%e re#ellion e3ist. ,n or(er to "re)ent it %a""ene(* %!$ans nee( to caref!ll& (esign t%e ro#ot s&ste$s. /ince %!$an create t%e A, to $ake t%e ro#ot $oral* .e (efinitel& can "re)ent t%e$ fro$ getting o!t of control #eca!se .e al.a&s t%eoreticall& s$arter t%an t%e A, .e (esign. Basic Programming for AI ,n t%e co$"!ter "rogra$$ing* $ac%ine cannot c%eck t%e error #& t%e$sel)es. T%e& 6!st si$"l& follo. t%e co$$an( .e gi)e t%e$. ,f t%ere are logical errors insi(e* t%e $ac%ine .o!l( not !n(erstan( .%at .e are asking t%e$. T%e& .ill 6!st follo. t%e e3act or(er .e gi)e t%e$. T%erefore* $aking a frien(l& A, is not si$"le. ,t re1!ires a .ise strateg& t%at .o!l( re1!ire (etaile( learning of %!$an )al!es an( "assing to t%e ne3t generation as a #asis for f!t!re nor$ali7ation an( reflection.+E* 8!(ko.sk&* "g 1Defining "Good" ,n or(er to "re)ent #a( conse1!ences of t%ose artificial agents* .e nee( to $ake consi(era#le effort on (efining .%at s%o!l( #e consi(er goo( .%en agents $aking (ecisions an( actions. 2o.e)er* #efore t%at* .%at (o .e $ean #& 9goo(9' :ne of (efinition is t%at 9goo(9 $eas!re against t%e s"ecific "!r"oses of (esigners an( !sers. ;!t* (efining .%at kin( of #e%a)ior is goo( for a ro#ot is not si$"le. 9,t is i$"ortant to (efer 1!estions a#o!t .%et%er a $ac%ine can #e gen!inel& et%ical or e)en gen!inel& a!tono$o!s 1!estions t%at t&"icall&
"res!$e t%at a gen!ine et%ical agent acts intentionall&.9+0.Allen "ara 2 <Superintelligence Mac%ine %a)e faster "rocessing s"ee(* t%e /!"erintelligence $a& #eco$e ac%ie)a#le. W%ic% $ean t%at t%ose $ac%ine can "otentiall& s$arter t%an %!$an. A &ear of t%inking can #e acco$"lis%e( #& t%e fastest $ac%ine in <1 secon(. ,t see$s s!"erintelligences co!l( (o e)er&t%ing. 2o.e)er* it .ill relate to t%e ca"a#ilit& an( $oti)e of t%e s!"erintelligence. 2ere are t%ree co$$on c%ain of reasoning regar( to t%e s!"erintelligence. A s!fficientl& "o.erf!l A, co!l( o)er.%el$ an& %!$an resistance an( .i"e o!t %!$anit&. T%erefore .e s%o!l( not #!il( A,. A s!fficientl& "o.erf!l A, co!l( (e)elo" ne. $e(ical tec%nologies ca"a#le of sa)ing $illions of %!$an li)es. T%erefore .e s%o!l( #!il( A,. :nce co$"!ters #eco$e c%ea" eno!g%* t%e )ast $a6orit& of 6o#s .ill #e "erfor$a#le #& A, $ore easil& t%en #& %!$ans. A s!fficientl& "o.erf!l A, .o!l( e)en #e #etter t%an !se at $at%. engineering* $!sic* art an( all t%e ot%er 6o#s .e consi(er $eaningf!l. T%!s after t%e in)ention of A,* %!$ans .ill %a)e not%ing to (o* an( .e .ill star)e or .atc% T= T%e 1!estion is t%at s!"erintelligence $a& not .ant to #!il( all of t%ose t%ing. 9T%e )ison lea"s (irectl& fro ca"a#ilit& to act!alit&* .it%o!t consi(ering t%e necessar& inter$e(iate of $oti)e.9+ E. 8!(ko.sk& "g<15Accor(ing to all of t%e infor$ation t%at , %a)e gat%er* $ac%ine et%ics is still a contro)ersial to"ic. ,t re1!ires so$e serio!s consi(eration. ;!t none t%e less t%at $ac%ine .ill not get o!t of o!r %an( like t%ose science fiction $o)ie s%o. !s. T%e
ke& "art is t%at %o. .e %an(le t%e co$"!ter !sage in o!r life an( %o. .e "!t $oral insi(e t%e s&ste$.
>eference Page
Yudkowsky, Eliezer. "Artificial intelligence as a positive and negative factor in global risk." Global catastrophic risks 1 (2
!"# $ $.

U4ab Wang Tao

Uploaded by

Copyright:

Available Formats

U4ab Wang Tao

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

U4ab Wang Tao

Uploaded by

Copyright:

Available Formats

Wang Tao Ms. Park ENGL 106 4/5/2014 Artificial intelligence A f!t!re "ro#le$ for %!

You might also like