Loading question...
You are in free guest mode. Timed Practice Revision List Attempt History Progress Tracking Create a free account to unlock these tools
Prev Next
Practice Reading: Fill in the Blanks (Dropdown) ID: #49820 AI 2
Instructions
A  survey  of  2700  AI  researchers  who  have  recently  published  work  at  six  of  the  top  AI  conferences  revealed  that  many  of  them  see  the  possible  future  development  of  superhuman  AI  as  having  a  non-trivial  chance  of  causing  human  extinction  or  other  extremely  bad  AI-related  outcomes.  Almost  58  per  cent  of  researchers  said  they  considered  that  there  is  a  5  per  cent  chance  of  such  .  The  survey  also  asked  participants  to  share  their  thoughts  on  possible  timelines  for  future  AI  technological  milestones,    the  good  or  bad  societal  consequences  of  those  achievements.  The  surveyed  researchers  predicted  that  within  the  next  decade,  AI  systems  have  a  50  per  cent  or  higher  chance  of  successfully    most  of  39  sample  tasks,  including  writing  new  songs  or  coding  an  entire  website  from  .  The  possible  development  of  AI  that  can  outperform  humans  on  every  task  was  given  50  per  cent    of  happening  by  2047,  whereas  the  possibility  of  all  human  jobs  becoming  fully  automatable  was  given  50  per  cent  chance  to  occur  by  2116.

Result:

With a free account: Retry and compare every attempt.
How difficult was this question?
Others found: 100% Easy · 0% Medium · 0% Hard
Saving...