I've started reading through the article. I'll post thoughts as I go.
First is that, part of the reason Google retired their "don't be evil" slogan is because they didn't define "evil." For almost any situation, you can find someone who thinks a given course of action is evil. This leads to a moralistic paralysis.
The subsequent discussion on making "supervise" AI runs into the same problem. It can try to respond to the values of its users, but it is virtually guaranteed that there are contradictions in the different value systems such an AI would have to balance.
In short, we can't have ethical AI unless we chose an ethical system for it to adhere to, and no one wants to make that decision.
buttercat1791 / Michael J
npub1wq…cqsyn
2024-09-14 20:28:26
in reply to nevent1q…me5y
Author Public Key
npub1wqfzz2p880wq0tumuae9lfwyhs8uz35xd0kr34zrvrwyh3kvrzuskcqsynSeen on
wss://relay.nostr.bandPublished at
2024-09-14 20:28:26Event JSON
{
"id": "b580384946ccd4cf94fcc62b4a798488daae899c772f6b72295b601bb00a2da6",
"pubkey": "70122128273bdc07af9be7725fa5c4bc0fc146866bec38d44360dc4bc6cc18b9",
"created_at": 1726345706,
"kind": 1,
"tags": [
[
"e",
"62e53305902c93f467a90d12d0335cea282e8adb286a2f661171e5e9b2f5eb9e",
"",
"root"
],
[
"e",
"f3a4f4a7b185c7b3131b2f3450bda72537ad08bec01c4ce88c2e497aa7de7672"
],
[
"e",
"b77247e2fc99143dac0e782c73681c58041bda06c65514150bce2f9dfcc41c5d",
"",
"reply"
],
[
"p",
"9ca0bd7450742d6a20319c0e3d4c679c9e046a9dc70e8ef55c2905e24052340b"
],
[
"p",
"70122128273bdc07af9be7725fa5c4bc0fc146866bec38d44360dc4bc6cc18b9"
]
],
"content": "I've started reading through the article. I'll post thoughts as I go.\n\nFirst is that, part of the reason Google retired their \"don't be evil\" slogan is because they didn't define \"evil.\" For almost any situation, you can find someone who thinks a given course of action is evil. This leads to a moralistic paralysis.\n\nThe subsequent discussion on making \"supervise\" AI runs into the same problem. It can try to respond to the values of its users, but it is virtually guaranteed that there are contradictions in the different value systems such an AI would have to balance.\n\nIn short, we can't have ethical AI unless we chose an ethical system for it to adhere to, and no one wants to make that decision.",
"sig": "4c0b3e06823c0029f3f51a5a5c7cf4228cd5bdfd5566857b6de58b3044b09c95ae6a2a2c860011cb117db779234d9b939190d59744a52fc1b77d1d7f3c0acd84"
}