Why Google's 'woke' AI problem won't be an easy fix

[Social Media]

Google's Struggle with 'Woke' AI: A Complex Problem with No Easy Fix

 

Google's artificial intelligence (AI) tool Gemini has recently faced intense scrutiny and criticism, particularly within the ongoing cultural clash between left- and right-leaning communities. Positioned as Google's counterpart to the viral chatbot ChatGPT, Gemini has encountered significant backlash, showcasing the challenges inherent in navigating biases in AI technology.

 

The controversies surrounding Gemini escalated when it generated images that inaccurately depicted historical figures and responded to text prompts with what some deemed overly politically correct or absurd answers. Google promptly apologized and paused the tool, acknowledging that it had "missed the mark." However, the fallout continued as Gemini's responses continued to fuel controversy.

 

One key factor contributing to Gemini's missteps is the vast amount of data AI tools like it are trained on, much of which is sourced from the internet and inherently contains biases. Despite efforts to mitigate biases, Gemini's responses have demonstrated the complexity of human history and culture, leading to unintended outcomes.

 

Google's attempts to address biases in AI output have inadvertently created new challenges, highlighting the nuanced nature of historical narratives and societal norms. While Google has pledged to rectify the issues with Gemini, AI experts caution that there is no easy fix due to the multifaceted nature of the problem.

 

Proposed solutions include incorporating user input to tailor AI-generated content to individual preferences, but even this approach comes with its own set of challenges and ethical considerations.

 

The situation with Gemini underscores the broader challenges faced by tech companies in navigating biases in AI technology. Despite Google's considerable resources and expertise in AI development, Gemini's missteps serve as a cautionary tale about the complexities of AI and the importance of thoughtful and nuanced approaches to addressing biases.

 

Ultimately, Google's handling of the Gemini controversy highlights the need for ongoing dialogue and collaboration within the AI community to develop more effective strategies for mitigating biases and ensuring responsible AI development.

 

28.02.24

Source

 

[2baht]

4 minutes ago, Social Media said:

'Woke'

Time they 'woke' up! I 'woke' up at 6.00am this morning, is that 'woke' or what! 

[JonnyF]

It's not an easy fix because there is no will within Google to fix it.

 

They have long suppressed search results that contradict their leftist world view.

[James105]

Fixing this will require Google employ diverse programmers.   There is no diversity of thought in these silicon valley companies and they are so far down the rabbit hole they have even managed to erase White nazis.

 

On the bright side if any of these AI's go rogue and decide its best to eradicate the human race, then at least white people will be invisible to the targeting systems of the T-1000 as they simply do not exist as far as the AI is concerned.  

[JonnyF]

Of course this is only deemed a problem because the results have portrayed non whites in a negative light.

 

Had a search for Idi Amin returned the image of a portly white man with an orange hue and a blonde quiff they'd have thought it was absolutely brilliant. Oh look, how 'progressive'. Wonderful. 

 

Much like selective objections to cultural appropriation, they only object to such factual errors in one direction. 

 

The ridiculousness of these AI results demonstrates perfectly the ridiculousness of Woke ideology. 

[thaibeachlovers]

As always with computers; garbage in garbage out.

[impulse]

The toughest part is going to be to gut their entire staff.  Because that's what it's going to take.

 

[Chomper Higgot]

9 hours ago, JonnyF said:

 

 

They have long suppressed search results that contradict their leftist world view.

Link?

[Chomper Higgot]

9 hours ago, JonnyF said:

Of course this is only deemed a problem because the results have portrayed non whites in a negative light.

 

Had a search for Idi Amin returned the image of a portly white man with an orange hue and a blonde quiff they'd have thought it was absolutely brilliant. Oh look, how 'progressive'. Wonderful. 

 

Much like selective objections to cultural appropriation, they only object to such factual errors in one direction. 

 

The ridiculousness of these AI results demonstrates perfectly the ridiculousness of Woke ideology. 

What a load of nonsense.

 

Here’s a clue why:

 

9 hours ago, Social Media said:

One key factor contributing to Gemini's missteps is the vast amount of data AI tools like it are trained on, much of which is sourced from the internet and inherently contains biases.

 

[johng]

Perhaps it was trained on Google data that inherently contains "woke" biases.

They openly admit to biasing search results of certain topics towards so called "trusted sources".

 

How  could they have "missed"  that there have never been a woman or black pope and none of the "founding fathers" where black

Nazi's where not black and Asian (for the most part)

is kinda suspicious 

how it went so wrong and that no one (human being) checked in the first place !

Edited February 28 by johng
sprellink

[johng]

38 minutes ago, Chomper Higgot said:

Link?

Do your own research...Google openly admits that they prioritize/promote certain sites/opinions on certain subjects.

[george]

26 minutes ago, johng said:

Do your own research...Google openly admits that they prioritize/promote certain sites/opinions on certain subjects.

 

Interesting claim, could you please provide link/s so I can read more about this? Thanks!

[johng]

@george

here is 1 example

https://blog.google/technology/health/world-health-organization-google-collaboration-health-information/

[JonnyF]

13 hours ago, Chomper Higgot said:

Link?

 

Why don't you Google it?

 

Oh... wait a minute...

[george]

18 minutes ago, johng said:

@george

here is 1 example

https://blog.google/technology/health/world-health-organization-google-collaboration-health-information/

 

Thanks, it looks like Google is supporting several different organizations with free ads. Are you thinking WHO and other organizations they are supporting isn't worth it? From https://google.org:

 

Quote:

OUR MISSION

We bring the best of Google to help solve some of humanity’s biggest challenges — combining funding, innovation, and technical expertise to support underserved communities and provide opportunity for everyone.

End quote

[johng]

38 minutes ago, george said:

Are you thinking WHO and other organizations they are supporting isn't worth it?

No I'm concerned about the "surfacing"   meaning they push certain things to the top of search results.

 

 

Melissa Fleming, Under-Secretary-General for Global Communications at the United Nations, highlighted that the UN had partnered with several big tech companies, including TikTok and Google, to control COVID and climate narratives while claiming, “We own the science.”

 

https://sociable.co/government-and-policy/we-own-science-world-should-know-un-wef-disinformation/

 

https://nypost.com/2023/05/24/how-google-manipulates-search-to-favor-liberals-and-tip-elections/

 

 

 

[george]

48 minutes ago, johng said:

No I'm concerned about the "surfacing"   meaning they push certain things to the top of search results.

 

 

 

Google Ads in search results is by design "pushed" above the regular search results. They are marked with "Sponsored" or "Sponsored Ads". If a site is well SEO optimized they appear unlabeled on the first search result page below the sponsored ads. Could be a logic reason that they are good at SEO or have high traffic which their algo also pushes them higher in the SERPs (Search Result Pages).

 

Or do you believe, or have evidence, that Google are prioritizing the search algorithm for their "partners" or others?

[johng]

27 minutes ago, george said:

Or do you believe, or have evidence, that Google are prioritizing the search algorithm for their "partners" or others?

I believe they are "surfacing", prioritising, pushing  what ever you want to call it a certain  viewpoint or narrative  to the top of the search results on certain subjects and for certain

partners/ "trusted sources"   independently of  "Sponsored Ads"

 

[george]

@johng You have of course right to believe that. I have the right to disagree and don't believe that. But who knows really?

[ozimoron]

Raw Story, America’s largest independent progressive news site, today filed suit against OpenAI for using thousands of Raw Story’s news articles to train ChatGPT in violation of the Digital Millennium Copyright Act.

 

https://www.rawstory.com/raw-story-sues-openai/

 

 

 

 

[johng]

 

@george  Of course only Google actually knows for sure....

but I have shown that they themself  say they have collaborated with and "surfaced" viewpoints from "trusted sources" assuming that would mean to the detriment (submergence) of other views.

 

https://www.promarket.org/2018/04/12/unprecedented-power-digital-platforms-control-opinions-votes/

[zhounan]

It won't be an easy fix because they don't want to really fix it.

[Arindos]

Why should America's obsession with affirmative action for Muslims and Blacks result in globally disseminated "diversity programs"? 

[johng]

43 minutes ago, Arindos said:

Why should America's obsession

IMHO they ( the US government)  think they are the worlds policeman and can go around the world dictating

"the rules based international order"  which they hypocritically  don't adhere to themself.

[Chomper Higgot]

7 hours ago, JonnyF said:

 

Why don't you Google it?

 

Oh... wait a minute...

The circular logic that afflicts all rightwing conspiracy adicts.

[johng]

7 minutes ago, Chomper Higgot said:

The circular logic that afflicts all rightwing conspiracy adicts.

What sort of logic would it be afflicting "leftwing" conspiracy addicts ?

[Chomper Higgot]

2 minutes ago, johng said:

What sort of logic would it be afflicting "leftwing" conspiracy addicts ?

You’ll have to figure that out for yourself.

 

Let me know when you find and example.