#ai #tech | Maria Sukhareva | 48 comments
Agree & Join LinkedIn By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Insights on Automation Bias in AI
That’s how you weaponise automation bias. (Automation bias = bias to trust machines.) I can’t judge those numbers and I genuinely don’t know which source it parrots - reliable or a random tweet. What I do know is that it’s very alarming when opinion leaders start citing stochastic parrots that were hyped up to AGI, Agents, and PhD-level intelligence to manipulate public opinion. Any tweet, article or news segment that quotes an LLM as the sole source of information, without knowing where the LLM got it from, should be discarded.
Concerns about Information Manipulation
Agreed! I see a dark pattern emerging here where LLM providers undercut other outlets of information, then they can dictate anything they want as fact. Same goes for posts dismissing other people's work used by millions as "stochastic parrots". The information in the grok response looks correct. As far as I can tell, the chatbots produce much more accurate responses than anything that has ever come out of EY, Deloitte, Mckinsey, KPMG, etc... and people... sorry, and midwits continue to trust those brands, so why not trust a chatbot in equal... sorry, so why not trust a chatbot in appropriately greater measure? Not too far from now, having access to data will no longer be the distinction, because there will be more data available than we can handle. Instead, being able to provide quality data people can trust, will be the value! The signal in all the noise!
Challenging the Status Quo
LinkedIn has never been a place for honest discussion but, well, bugger it, let's say this openly: one can hardly expect any words of meaning or wisdom from a delusional pro-nazi billionaire. I think the best we can do is not to quote such personas and give their words undue significance. It looks like it (Grok) is actually making a prediction: US (national) deficit is a bit short of $1.05 trillion, for fiscal year 2025. For context debt is short of $37 trillion or around 123% of GDP based on the FRED data for H1 2025.
Ensuring Quality Data and Trust
The Problem with this "opinionated content" peddling is that it deprives people of the variety of information and viewpoints (even what's on the first page of Google ), and you can always tweak the model to weigh (bias) one narrative or another... How does that differ from quoting LLM for any use? Agree but based on the current misinformation we are facing, I trust more LLMs than fake journalists.
To view or add a comment, sign in
14,515 followers Create your free account or sign in to continue your search
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
