Here's the result from the brainstorm session last night, with some
comments added.

[...]

First we have proplsed replacement terms that got >= 7 votes.  The
columns are a "*" if we gave people a chance to speak for or against
the term, and the number of votes.  Each person got to cast 1 or 0
vote for each word, independently.  There were at least 12 people
present, since "Safe AI" got 11 votes and I didn't vote for it.

       9       Friendly AI
       11      Safe AI
       8       Safely scalable AI
       7       Humane AI
*       8       Benevolent AI
*       9       Beneficial AI
*       7       Safe at any speed AI

Then we have the terms that got < 7 votes.  The first column is the
number of votes.  One person cast a very lukewarm vote, hence the 2.5.

4       Humanistic AI
0       Non-lethal AI
0       Genie AI
3       Stably benign AI
2.5     Safe and beneficial AI
4       Stably safe AI
3       Stably humanistic AI
0       Pleasant AI
2       Good AI
2       Safe Seed AI
0       Nice AI
1       Do what we mean AI
0       Do what we want AI
0       Don't do what we don't want AI
0       Don't do what we don't mean AI
5       Moral AI
1       Non-catastrophic AI
0       Predictably non-catastrophic AI

Then we have terms that were proposed or written down only after the
main voting, so we didn't vote on them at all:

       Human-aligned AI
       Phlegmatic AI
       Cautious AI
       Careful AI
       V
       Scalable AI
       Scalable safe AI
       Incorruptable AI
       Dynamically stable AI
       Intrinsically safe AI
       Dynamically safe AI
       Invariantly friendly
       Wise AI
       Artificial Wisdom
       Mostly-harmless AI
       Rarely chaotic evil AI
       Lawful good AI
       Neutral good AI
       Safer AI

[...]

May we add those if they are fit enough ?


Unerringly Humane AI (UHAI) (alt:Undeviably Humane AI)

Steadfast AI (SAI)

Ethically discerning AI (EDAI)

Loyally Humane AI (LHAI)

Artificial Moral Intelligence (AMI) (as in french "ami" for "friend")

Obdurately Honorable Synthetic Humane Intelligence (OH SHI) (what we'll think if we don't get it right)