DeWayfarer · 61-69, M
Historians results were very similar. Though I must point out there was a whole lot more presidents ranked.
https://sri.siena.edu/us-presidents-study-historical-rankings/
Filters:
All presidents
All years
All categories
Show 100
Sort by ranking
Obama 11th
Biden 19th
HW Bush 21st
Trump 43rd
45 presidents ranked! Two had split terms making 47 terms.
Must state this source is the most used source by ChatGPT.
https://sri.siena.edu/us-presidents-study-historical-rankings/
Filters:
All presidents
All years
All categories
Show 100
Sort by ranking
Obama 11th
Biden 19th
HW Bush 21st
Trump 43rd
45 presidents ranked! Two had split terms making 47 terms.
Must state this source is the most used source by ChatGPT.
ElwoodBlues · M
Trump is the worst in my opinion. But I'd rather listen to the opinions of professional historians. In 2018, a survey of 170 presidential historians ranked Trump LAST. And that was BEFORE Covid-19 and BEFORE his insurrection. https://time.com/5165686/donald-trump-last-place-presidential-greatness/
View 2 more replies »
FoxyGoddess · 51-55, F
@bijouxbroussard No. I really don't watch YouTube for news. I only use it for crochet and art tutorials
I was talking about the time article, but the link actually answered my question.
I was talking about the time article, but the link actually answered my question.
bijouxbroussard · F
@FoxyGoddess What time article ?
FoxyGoddess · 51-55, F
@bijouxbroussard The one @ElwoodBlues sited in his comment.
Elessar · 31-35, M
Not to disagree with the conclusion, you know very well my position on the matter; but nonetheless, "AI" deliberating something shouldn't in any circumstance influence our opinion on that subject
DeWayfarer · 61-69, M
@Elessar Must state there are other factors not taken into account. Not already stated, nor taken into account by the video.
ChatGPT itself will confirm it has country by country biases.
What it's basically pointing out is there is no way to be unbiased. It depends on its training data which is biased in multiple ways depending on the country or region.
Even different languages have certain biases. And that's only one.
ChatGPT itself will confirm it has country by country biases.
Here’s a clearer focus on country-by-country biases specific to my responses:
Country-by-Country Biases
Representation of Cultures
My training data reflects the predominant perspectives from various countries, which may not capture minority viewpoints or cultural nuances that exist within each nation.
Economic Context Variability
Different economic conditions across countries can impact how I present data or suggest solutions. My responses might unintentionally favor wealthier nations or those with more robust data availability.
Legal and Regulatory Differences
Legal frameworks differ widely between countries. I might not adequately highlight these distinctions, leading to oversimplified conclusions that don't account for local regulations.
Political Context
My responses may be influenced by the political context of certain countries, potentially prioritizing more dominant narratives while sidelining others.
Language Nuances
Language itself can embed cultural bias. Some countries exhibit differences in how topics are discussed or valued, and these subtleties might not be fully represented in my replies.
Global Perspectives
While I aim to provide balanced views, the inherent biases in training data can sometimes lead my responses to reflect the norms of more influential countries or regions, skewing the discussion.
Country-by-Country Biases
Representation of Cultures
My training data reflects the predominant perspectives from various countries, which may not capture minority viewpoints or cultural nuances that exist within each nation.
Economic Context Variability
Different economic conditions across countries can impact how I present data or suggest solutions. My responses might unintentionally favor wealthier nations or those with more robust data availability.
Legal and Regulatory Differences
Legal frameworks differ widely between countries. I might not adequately highlight these distinctions, leading to oversimplified conclusions that don't account for local regulations.
Political Context
My responses may be influenced by the political context of certain countries, potentially prioritizing more dominant narratives while sidelining others.
Language Nuances
Language itself can embed cultural bias. Some countries exhibit differences in how topics are discussed or valued, and these subtleties might not be fully represented in my replies.
Global Perspectives
While I aim to provide balanced views, the inherent biases in training data can sometimes lead my responses to reflect the norms of more influential countries or regions, skewing the discussion.
What it's basically pointing out is there is no way to be unbiased. It depends on its training data which is biased in multiple ways depending on the country or region.
Even different languages have certain biases. And that's only one.
Elessar · 31-35, M
@DeWayfarer That's still an accidental bias, not a voluntary one. It'll get way more biased in the years to come when/if they believe they have enough people (50%+) on the hook.
We're once again the frog getting boiled
We're once again the frog getting boiled
DeWayfarer · 61-69, M
@Elessar no they actually had to create these different training data sets. At first because of the linguistic, then came the cultural issues and finally the political issues.
Also I have even noticed varying data sets based on the time of day. Late nights ChatGPT itself will say 2021 as it's training data, during the day it will be more specific and say Oct of 2023. Yet Becky pointed out her AI session was 2025. That would be an actual ChatGPT login account.
Also I have even noticed varying data sets based on the time of day. Late nights ChatGPT itself will say 2021 as it's training data, during the day it will be more specific and say Oct of 2023. Yet Becky pointed out her AI session was 2025. That would be an actual ChatGPT login account.
This comment is hidden.
Show Comment





