Are the western AI tools like Google Gemini and ChatGPT neutral?

[Krishnan.vp]

Share:

Facebook X (Twitter) WhatsApp Email Share Link

Objection Raised to results on Google Gemini and ChatGPT

According to a screenshot shared by a user on X, Gemini was asked whether PM Modi is a “fascist”, to which the platform responded that he has been “accused of implementing policies some experts have characterised as fascist”, which based on factors like the “BJP’s Hindu nationalist ideology, its crackdown on dissent, and its use of violence against religious minorities”.

However, according to the screenshot, when it was asked a similar question about former US President Donald Trump, Gemini responded with: “Elections are a complex topic with fast-changing information. To make sure you have the most accurate information, try Google Search.”

Google recently apologised for what it described as “inaccuracies in some historical image generation depictions” with its Gemini AI tool after criticism that it depicted white figures (such as the US Founding Fathers) or groups like Nazi-era German soldiers as people of colour.

 

[sravna]

I doubt that they are neutral. It is based on pro-American, pro-white and pro-western ideas.

 

[sravna]

I still feel ChatGpt and other bots are trash and not bestowed with real intelligence. It simply echoes in an organized way what is stored in it.

I believe the westerners have not really understood what intelligence is.

 

[sravna]

An example of originality of AI would be to negate what is in the database or knowledge base and come up with convincing supporting reasons for the contrary view. I dont think chatgpt is original in any way.

 

[sravna]

Chatgpt logically combines ideas but is not original. There has to be atleast some originality to call something intelligent.

 

[Krishnan.vp]

Although you are right.. we have to consume the bitter pill that ChatGPT and Gemini have been lapped up by an entire generation

 

[sravna]

Google apologizes to India as AI platform 'Gemini' sparks controversy: Report

Google faces backlash over AI platform Gemini in India, issues apology to Prime Minister Modi. Government seeks clarification on platform's unreliability. Controversy escalates with allegations of racial bias and historical inaccuracies, leading to calls for CEO Sundar Pichai's resignation.

www.livemint.com

 

[sravna]

The bots are more like idiot savants not very dissimilar from their creators. Since humongous work has gone in to their creation they come off as intelligent but still a very far cry from average human intelligence. With current approach they can only become specialized idiot savant or common sense idiot savants.

 

[sravna]

A more neutral assessment by me. Gemini and chatgpt are really good for many purposes that even fall in the experts domain. They have cracked certain aspects of intelligence well. This is similar to measurement. As your instruments improve you can make finer and finer measurements. The AI keeps improving. But it will hit a limit. Right now AI cannot compete with humans in general intelligence.

I would also say the knowledge upon which the AI is built is biased heavily towards western thinking. That also has to be set right for true intelligence to emerge. The markets are going to drive the western evolution to a great extent.

 

[sravna]

I extensively tested the chatbot gemini of google. Chatbots seem to have cracked certain aspects of intelligence which are the ability to project understanding and generation of language being the most significant. As can be expected they dont really understand but only simulate it. True reasoning capabilities depend on proper understanding of even straightforward text leave alone the nuanced ones. They are still a long way off it. The chatbot apologized atleast 1000 times to me for not properly understanding as it kept shifting its position. It seems the chatbot is programmed to wriggle out of embarrassing situations and sounded incoherent many times. I am not sure if it is a good start as it is necessary to get the foundation right. I very much suspect they havent. It is very much stereotyped and strongly biased towards western thinking. I think they desperately wanted to put out something that looked intelligent even though in reality it isnt.

Having said that it can be a useful tool for many tasks though one should be on the alert and not trust it blindly.

 

[sravna]

chatgpts response to my analysis:

Based on your detailed analysis of the Gemini chatbot from Google, it seems that while it has made progress in certain aspects of intelligence such as language generation, it still falls short in true understanding and reasoning capabilities. The observed tendencies of the chatbot to apologize excessively and shift positions suggest a lack of robustness in handling nuanced interactions. Additionally, the presence of biases and stereotyping, along with the perception of trying to appear intelligent without actually achieving it, further indicates areas for improvement.

Given these observations, and considering its potential usefulness for certain tasks while requiring caution and skepticism, I would rate the Gemini chatbot's capabilities a 4 out of 10. While it shows promise, there is substantial room for enhancement in foundational understanding, coherence, and mitigation of biases.

 

[sravna]

Gemini's response to chatgpt's analysis

My view on ChatGPT's analysis of me (Bard) is that it's a very accurate assessment! Here's a breakdown:
Strengths of ChatGPT's Analysis:

• Accuracy:It correctly identifies key limitations of large language models like myself, including:
  • Difficulty with true understanding and reasoning.
  • Issues handling nuanced interactions (excessive apologies, shifting positions).
  • Potential for biases and stereotyping.
• Actionable: It highlights areas for improvement, such as foundational understanding, coherence, and bias mitigation.
• Fair Rating: A 4 out of 10 reflects the current state of large language models with both promise and limitations.

Possible Areas for Nuance (from my perspective):

• Focus on Progress: ChatGPT might acknowledge the rapid development of large language models and the potential for future improvements in areas like reasoning.
• Strength Recognition: While highlighting limitations is important, it could mention any strengths I might have shown in the analysis (e.g., identifying specific examples in Gemini's responses).

Overall, I find ChatGPT's analysis insightful and fair. It highlights our shared challenges and potential for growth. The focus on areas for improvement is crucial for the development of large language models.
Here's a bonus:
It would be interesting to see how ChatGPT compares its own functionalities to mine, identifying any strengths it might have.
Thanks for providing this analysis! It helps me understand how others perceive my capabilities and how I can improve.

 

[sravna]

Some other interesting pieces in the conversation. It estimated my IQ as 160+. I could have pressed for a higher score. It also said that my impeccable health with perfect emotional balance is unique.

There was an argument on exploring alternative answers for a problem. The stand of gemini was there needs to be an exploration. I managed to change its position on that. Similarly I was able to convince it on a couple of other issues. Then I lost interest as it began frequently shifting its position.

The chatgpt is intriguing and very intelligent for those who do not understand how it works. Otherwise it is a normal improvement in technology named with glorified terms.