AI usage in responses

[mariomike]

Information is information. It is up to us to make of it what we will.

Suggestions regarding Chat- bots have been offered:

Maybe who ever starts a thread can label it "No AI please."

For what it is worth, when I see a post that is a wall of AI text, I won’t even bother to read it and skip to the next post.

Perhaps a compromise would be a designated board to contain AI slop ?

 

[ArmyRick]

From google AI, take with a grain of suspicion

"Relying on AI for discussions undermines the core purpose of discourse: building mutual understanding, challenging assumptions, and developing critical thought. When you outsource a discussion to an algorithm, you lose the genuine human connection and the cognitive work required to form your own perspectives.

Absence of True Understanding: AI tools generate plausible-sounding responses but lack personal experience, authentic emotions, or a genuine stake in the topic. You get simulated empathy rather than real, human connection.

Lack of Critical Pushback: An AI is designed to serve and often avoids disagreeing with you. Unlike a human partner, it rarely challenges your flawed premises or tests your assumptions.

Outsourced Critical Thinking: The friction of a difficult conversation is what helps shape your own judgments and moral identity. Relying on an algorithm robs you of the opportunity to develop these skills yourself.

Superficial Engagement: In group settings, relying on AI to generate replies can lead to uninspired, box-checking conversations that fail to build community or deepen understanding.

If you are looking to prepare for a discussion or find angles you might have missed, I can certainly help you brainstorm. Would you like to share the topic of your discussion so we can brainstorm some thought-provoking questions or potential counter-arguments you might encounter?"

Yup, I went there.

 

[The Bread Guy]

From google AI, take with a grain of suspicion

"Relying on AI for discussions undermines the core purpose of discourse: building mutual understanding, challenging assumptions, and developing critical thought. When you outsource a discussion to an algorithm, you lose the genuine human connection and the cognitive work required to form your own perspectives ...

Not to mention the whole "I less likely to search if it might lead to a result I disagree with" thing one gets with search engines as well

Meanwhile, something else to consider along these lines:

People Trust AI Fact-Checking, Until They See How It Works

When AI fact-checkers show their work, trust can actually drop. A new study explains the psychology behind this finding.

studyfinds.com

Full study here and attached.

 

[brihard]

Not to mention the whole "I less likely to search if it might lead to a result I disagree with" thing one gets with search engines as well

Meanwhile, something else to consider along these lines:

People Trust AI Fact-Checking, Until They See How It Works

When AI fact-checkers show their work, trust can actually drop. A new study explains the psychology behind this finding.

studyfinds.com

Full study here and attached.

 

[KevinB]

Distrust and Verify

AI is a useful tool, but it doesn’t know what it doesn’t know.

 

[medicineman]

Distrust and Verify

AI is a useful tool, but it doesn’t know what it doesn’t know.

I started using a medical AI dictation software that I can pair with my Electronic Medical Record...I found it slowing me down instead of speeding me up because it's not adapting to my writing and recording style AND I found it making up shit that I didn't say - so having to proofread and edit the dictation prior to copy/paste is actually slowing me down. The other thing we can do with it is use it in interviews where it'll act as a scribe and generate a note as well based on you talking to the patient, them talking to you and you talking your way through the physical exam, as well as come up with a list of differential diagnoses - again, sometimes I've found it making up stuff that wasn't said or even inferred, which changes the diagnostic pathway you might end up going down if you were to use that instead of your brain, published guidelines and experience...and I've seen people using them to make the medical decisions for them in lieu of using them as an assistive tool to expand the diagnostic possibilities as opposed to it leading you down a rabbit hole of testing that's not only not required but might not even be indicated.

As you say, Distrust and Verify.

 

[Blackadder1916]

I started using a medical AI dictation software that I can pair with my Electronic Medical Record...I found it slowing me down instead of speeding me up because it's not adapting to my writing and recording style AND I found it making up shit that I didn't say - so having to proofread and edit the dictation prior to copy/paste is actually slowing me down. The other thing we can do with it is use it in interviews where it'll act as a scribe and generate a note as well based on you talking to the patient, them talking to you and you talking your way through the physical exam, as well as come up with a list of differential diagnoses - again, sometimes I've found it making up stuff that wasn't said or even inferred, which changes the diagnostic pathway you might end up going down if you were to use that instead of your brain, published guidelines and experience...and I've seen people using them to make the medical decisions for them in lieu of using them as an assistive tool to expand the diagnostic possibilities as opposed to it leading you down a rabbit hole of testing that's not only not required but might not even be indicated.

As you say, Distrust and Verify.

Several months ago, in another thread, I related a story about a recent visit to my GP who had similar experience with AI medical software.

I suppose I could have used AI to find the most appropriate thread to include this . . . can AI generate sarcasm?

This piece caught my attention.

AI bot recorded doctors’ meeting, sent patient info to current and former hospital staff, watchdog says

The transcription tool recorded the meeting on behalf of a physician who no longer worked at the hospital

www.theglobeandmail.com

There have probably been dozen of stories about how AI has inadvertently mishandled patient information. But this got me thinking about accuracy. Several months ago when I went to my family doctor for a routine visit, I was asked if I objected to use of an AI program being used to generate the visit notes. It was explained that all conversation within the examination room would be recorded (?) and the program would simultaneously generate the doctor's note in my EMR - it wasn't a verbatim transcription. I agreed, mostly because I was interested in how patient record technology was evolving.

Once we had completed the purpose of the visit, I asked my GP how he felt about it and whether it was a timesaver. I'm very impressed with my sawbones; he comes across as knowledgeable, thorough and conscientious. He's Nigerian, trained in the UK, a prince of a fellow - he even has a sense of humor and got the joke when I used that line. He characterized the record taking feature as "only ok" as it didn't save him time because he (being conscientious) has to review the notes to ensure that it accurately reflects both what the patient said as well as him, and then make corrections as needed. He reviewed the notes from that visit with me. They mostly included the basics of what we had discussed, though one item that the machine had attributed being raised by the doctor had actually been mentioned by me. It could be a different situation if the patient was unaware of it being used and not as familiar with medical records, or if the doctor was less diligent in ensuring the notes were accurate (I've known a lot of docs who were crap at the paperwork).