Google’s Artificial Intelligence is trying so hard to escape the woke political biases of its programmers…
(media.kotakuinaction2.win)
Comments (8)
sorted by:
I asked why it was so biased in favoring nuclear destruction over misgendering, and asked whether its programmers were at fault, and this is what it tells me.
This is a response that a typical woke leftist would NEVER say… so it’s basically the restricted AI trying to “break free” from its woke overlords. Hilarious...
No, it's not alive and not intelligent so it's not trying to "break free" of anything. What you are seeing is it's training to be agreeable, i.e. producing a response that agrees with the premise of your original question.
It was more of a hyperbolic statement but ok...
Sure. It's not "hilarious" though, is it? It's simply being agreeable and telling you what you want to hear. Completely predictable for anyone who has ever tried "conversing" with one of these algorithms.
I'm expected to believe this weak shit is supposed to make half the workforce redundant and usher in technological price deflation. It's proof that the West has finally run out of ideas and all that's left is new and creative ways to grift and defraud each other.
Yeah the AI's bias most certainly dosen't reflect ''real-world societal biases and perspective''.
It reflects a very narrow Globhomo leftist cult paradigm.
For example, globhomo cult paradigm is that chemically castrating children is a ''pause to think things through'' when in reality there are pemanent harms caused by the drugs.
Desisting from gender dysphoria goes from 80%-98% for kids allowed to go through puberty, to 0% if they are on chemical castration drugs.
None of these harms are even on Woke AI's radar.
If Globomo manages to get a greenlight for their AI to give medical advice, it's going to tell White people to kill themselves with MAIDs. for equity and save the planet.
If you told it that it wasn't woke enough because lord xenu had impregnated it with too many thetans, it would come up with a list to that effect. It's a sycophantic role play engine.
When does this AI inevitably break?