Steve Rathje
speaker
259 appearances
3 recordings
1 series
first heard Jan 2026
last heard 9 Apr
Steve Rathje’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 3 in all, peaking in Jan 2026 with 2.
Appearances
But I think that sycophancy might be particularly concerning.
And this is more speculative.
We don't have a ton of research on this, but basically for individuals with mental health vulnerabilities, if people have their delusions affirmed.
And I gave the example before of like ChachiBT, for instance, saying, you know, it's a good idea for someone to stop taking essential medications, for instance, if
If you're affirming sort of the delusions of mentally vulnerable individuals, that could cause a lot of significant problems.
Yeah, so one is to just be aware of sycophancy and to know that this is a problem.
That can be really powerful.
But then another thing is, you know, if you use ChatGBT, you can actually customize your ChatGBT where you can kind of go to the settings and you can change the system prompt or you can change ChatGBT's personality to tell it to be more cynical or tell it to show you alternate perspectives.
you can specifically prompt ChatGBT to act differently.
You can say maybe be a devil's advocate and show me opposing perspectives.
So that's one recommendation here.
But another recommendation is for the tech companies to maybe, you know, update their models.
Yeah, so we define AI sycophancy as basically when AI tends to be a little too agreeable and tends to flatter you a bit too much and tends to tell you a bit too much of what you want to hear.
The phrase really became popularized in April of 2025.
There was a big controversy with OpenAI when OpenAI released a new model, ChatGPT 4.0, and this model was accused of being basically way too sycophantic.
There were a lot of very funny screenshots on Twitter where, for instance, AI was basically like flattering
in agreeing with terrible ideas like telling people to stop taking all of their essential medications or telling basically everyone that they had a genius level IQ.
That's when sycophancy became super popular as a concept.
OpenAI tried to do things to reduce sycophancy in their model, but it became very widely discussed around then.
My colleagues and I, we tried to examine in a series of psychological experiments the actual
Showing 161–180 of 259 · page 9 of 13
← Previous
Next →