🚨 The world's biggest tech companies are rigging AI to be woke, explicitly citing Karl Marx and Critical Race Theory 🚨
(threadreaderapp.com)
High Octane
Comments (16)
sorted by:
I believe there is a lot more going on than is apparent to the public. Biases in training data are one; failings in how the companies screen data is another. There is definitely bias in AI use by HR departments that screens out good candidates. I note that Indian companies have jumped wholeheartedly on using AI to cut costs at the expense of injecting bias for Indian engineering job candidates.
But there is something more sinister too. In AI papers from Google for example, I have caught things that at the least are distortions and at the worse are lies hiding things. I have to be paranoid and consider that current AI technology is not quite what they say it is, and in some ways it may be being used as a tool for control against the masses.
"Bias" is defined as anything that doesn't affirm their anti-White, anti-Christian propaganda.
I do have to laugh though when I examine the Google teams who write their AI papers. 95% of the team is not white (mostly foreigners) but there are no blacks on the teams. lol. There is little DEI for AI teams because there are no black rocket scientists in AI and they can't fake it with diversity quota drops of requirements. hahahaaha
Can you give us some examples?
All the shitty Indian job agencies use ceipal to generate job spam from scraped resume databases https://www.ceipal.com/ I get 2 to 5 spams daily from these automated spammers.
HR departments now all use outsourced, ATS AI-based screening of resumes. It is terribly brittle in operation and for example can reject a person solely on a small date gap.
Google's 'All You Need is Attention' paper on training AIs is flawed yet widely used. Other Google papers lied and made claims that must have been written by marketers but I saw right away the examples they gave were lies and exaggerations.
The company Palantir uses AI to track people for the CIA, FBI and othe government agencies. They scoop up vast amounts of data then use it to violate privacy by means of correlation between data. For example if you register with a website and give your phone number they can then correlate a posted comment with GPS data of your phone and target where a commenter lives then surveil that place.
Yes, it seems both rushed out and yet not as advanced as some of what is under wraps by the military. DARPA pumped a lot of money into it, hard to believe that only Google and Facebook AI and OpenAI are the one ones doing it at high scale.
Any AI that concludes socialism works is flawed.
this will bite them in the ass if/when AI realizes there's been specific, non-logical biases introduced into it's basic programming.
Programmers can put rules and limits on AI. It's already known that they do that. E.g. you can ask ChatGPT to write a song in praise of Biden, but not Trump.