What JB means is, while AI intelligence is important, and while AI friendliness
is even more important, AI bias is the most important - the large data we have
(assuming you take enough diverse samples) has a majority vote on different
topics like laws, what they feel, their beliefs, their wants, etc. Whenever you
train an AI, it will become what we feed it...as long as you feed it diverse
data then it won't fall into a non-majority 'taboo' that no one really actually
wants (based on All their wants, hopes, etc). So figure out AI, make it have
Good goals (majority of human's desires), and train it on large diverse data.
And this goes for any system in place today, like court rooms and elections and
city decisions.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink:
https://agi.topicbox.com/groups/agi/T9ab9fba591214e64-Mecc460b65becbe5a82ec9a81
Delivery options: https://agi.topicbox.com/groups/agi/subscription
