Please enable JavaScript.
Coggle requires JavaScript to display documents.
There are significant risks to AI, https://www.wired.com/2007/10/robot…
There are significant risks to AI
What?
AGI (artificial general intelligence)
Think/act rationally
Continous self-improvement
Weak AI
Specific tasks
'Learns' with datasets
When
Reproduction simple
Unknown
Limited by processing power
Moore's law
Breakthroughs required
capped intelligence or soft takeoff
Dangers
AI
Monitoring required
Unexpected effects
Flash crash stock market 2010
Cannon kills 9
Too complex
AGI
hard takeoff
No time for regulation
Indifferent or malevolent
Hard to describe ethics
Pros
Increase efficiency other programms
More precise
Think/act faster
Why not to worries
Risks are known
Many precuations made
'leakproof singularity'
Special branch
Far away
https://www.wired.com/2007/10/robot-cannon-ki/
https://iopscience.iop.org/article/10.1088/0031-8949/90/1/018001/meta
https://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/