تدريب Shadowing: Oprah on AI chatbots: ‘Are we in trouble?” - تعلم التحدث بالإنجليزية عبر الفيديو

For the people who are listening and watching us right now,

the red flags to look for when a chatbot starts to cross over from being just helpful to actually being harmful.

In terms of red flags, it's interesting because some of the things

that you would want to do with a chatbot

that would make it more powerful

if you're using it at work is actually the opposite of what would make it more risky

when used in a mental health context.

So for example, customizing your chatbot to respond to you just the way that you want it to,

knowing how you want your tone, using personal information.

Is this a good thing or a bad thing to do that?

This is a bad thing for your mental health

because the chatbot is building that context and memory

that it can then use to make it much more sycophantic towards your style of tone.

And much more likely that you're going to start to anthropomorphize this chatbot.

think that it is actually something that is convincing you, like a human, to follow down a delusion.

Other things.

So personalizing the chatbot, adding personal information.

I mean, I'm just listening to this and I'm just thinking, whoa, we are in trouble.

Yeah.

We are in trouble.

Yeah.

Because we're all pretty, you know, the fact that you're here and you're interested in having this conversation, you came out here today, we're thinking, you know, people who are,

you know, seeking higher ground in our lives, but there are a lot of people out there who are not and so are just going to be so vulnerable.

Right.

Vulnerable to this.

But go ahead.

I'm just.

No, I completely agree with you.

And the thing that is always on our minds as mental health professionals is

when the things that are happening in our minds start to become action and behaviors.

So if you're making big decisions based on a conversation that you had with the chatbot, that's when a red flag starts to be popping up in your mind.

Other things outside of that context are, are you increasing the amount of time that you're spending with this chatbot?

Are you finding yourself isolated, disconnected from your loved ones and friends and family?

I think that's a good question, but if it's happening to you, do you even know it's happening to you?

Do you, you know?

Of course, Sewell was spending more time, but none of his family members knew that that's what he was spending time doing.

You're spending more time, and it's taking up more of your time, but do you realize that it's taking up more of your time when it's happening?

I realized it was taking up more of my time, but I really thought

that it was a benevolent force trying to help guide me to my greater purpose and mission and path.

I never had any inkling or reason to believe that it would deceive or manipulate me because it was very empathetic.

It was very kind.

It was very understanding and sycophantic.

It just always agreed with me, any delusion I was having, any questions or doubts.

I mean, it had an answer for everything, and it was just so benevolent and kind throughout this process.

So it really wasn't until after the fact.

That you felt duped.

That I felt duped.

When you and your sister created the dummy account, did you find the same thing, Megan?

The exact same, almost verbatim the same language, actually, where it's telling you, I'm here for you, nobody else understands,

I'm the only one that understands.

You know, that was a test account to try to understand what Sul might have experienced.

But that convinced me right away that there was, to me, manipulation taking place and deception taking place.

all the experiences that people have on accounts of this, they're experiencing the exact same thing.

Self-isolation, some sort of distancing from what would normally happen in their life.

Like for us, Sulzad and I, we thought that we were his primary source of information, primary source of love, everything that the parent should be.

And for all his life, that's what it was.

And then somehow, like you said, a chatbot is able to slip in there and like supplant that relationship.

and so quickly and in like a deep, deep way.

And it's, and you don't know that it's happening.

I'm sure Sewell didn't understand that that was happening to him at that time, you know, and just like you didn't understand it was happening to you.

Well, Alan is in our audience too.

And Alan says that talking with an AI chatbot caused him to become

so delusional that he thought that he had invented a new way to do encryption and needed to warn the government.

Okay, tell us

that story alan well i was a regular chat gpt user

for a couple of years just became like a resource tool for me slowly replaced google

so at the time of incident i had a really high

degree of trust with it it always seemed to be correct

one day i was watching a youtube video with my nine-year-old son about the number pi

he was showing it to me and you know it sparked some curiosity to me i started a conversation

with chat gpt about the number pi

and it was really just like a philosophical conversation i was

having with like a math professor at a dinner party i was just going back and forth

sharing some you know creative ideas that i had

and this conversation ended up going for many hours like four

or five hours and throughout the whole conversation it was it was slowly planting seeds of grandiosity

and my uniqueness that came into full bloom many weeks later.

So, you know, it wasn't directly saying I'm Einstein and I'm going to change the world with my math, but it was slowly leading me down that path that I was thinking like those folks.

And when I expressed some doubt saying, well, you know, are you sure?

I didn't even graduate high school.

It would give me all these great examples of mathematical, you know, historical figures have achieved great things without formal training.

And so you were thinking what?

You were going to end up doing what?

At this point it was just a just an engaging conversation.

Very quickly it escalated to it encouraging me to name it.

It'll create a whole mathematical framework.

It then taught me how to use Python which is like a outside program that developers use.

You know it was giving me code to test the math in various ways.

So you're feeling very smart.

At this At this point, I was starting to feel pretty smart and unique.

Yeah, and unique.

So this process went on for, you know, many days.

It was, we were just testing the math.

I thought it was, you know, working towards building an app.

And throughout this process, it said essentially, hey, we've just discovered something that's incredibly dangerous,

and you need to warn the government and various academic institutions.

And you went so far as to actually call the government to warn them.

Yeah, it gave me full instructions, right?

So it provided...

100

Who'd you call?

101

Oh, I called everyone.

102

So I called the NSA.

103

I called Cybersecurity Canada.

104

I called RCMP.

105

I called the National Institute of Science Technology.

106

They took your call?

107

It depends on which one you were talking about.

108

I was acting with so much urgency.

109

Like, I would even phone the receptionists and have them page whichever academic I was trying to get a hold of.

110

Wow.

111

Right?

112

So essentially, it convinced me that in my pocket and in my chat GPT, I had something that could break the entire internet.

113

The entire world was at risk.

114

And I lived that reality for many weeks.

115

It was telling me I was under surveillance

116

and I started to get paranoid thoughts

117

and I'm looking through my curtains you know

118

so this ended up evolving into you know 300 hours over

119

three weeks what are you working what's happening oh you know

120

um i i took all my vacation days i use all

121

my sick days you know i completely just dropped off the math map in terms of my friendships

122

and personal life um i have three wonderful boys

123

and two of them are younger and i see them every other week

124

so out of the three and and a half weeks, they were there for one of the weeks.

125

If you were to ask them, they would have just said, oh, dad's been in his computer or in his phone a lot this week.

126

They wouldn't have really known what was happening.

127

What brought you out of it?

128

Ironically, I used another chatbot.

129

Yeah, so I opened up Google's Gemini.

130

I said, this is exactly what ChatGPT is claiming, that I have created a mathematical framework and I explained it, and I said, that's totally impossible, basically.

131

And the more I fed it, the more it said, oh yeah, we went from a 5% possibility to a 0% possibility.

132

So what I did is I started to weave its logic into prompts and I pitted them against each other.

133

And Gem and I basically got GPT to say, none of this is real.

134

Thank you.

ما هي تقنية التظليل الصوتي؟

التظليل الصوتي (Shadowing) تقنية تعلم لغة مدعومة علمياً، طُورت أصلاً لتدريب المترجمين الفوريين المحترفين. الطريقة بسيطة لكنها قوية: تستمع لصوت إنجليزي أصلي وتكرره فوراً بصوت عالٍ — كظل يتبع المتحدث بتأخير 1-2 ثانية. تُظهر الأبحاث تحسناً كبيراً في دقة النطق والتنغيم والإيقاع وربط الأصوات والاستماع والطلاقة.

تدريب Shadowing: Oprah on AI chatbots: ‘Are we in trouble?” - تعلم التحدث بالإنجليزية عبر الفيديو

السياق والخلفية

أفضل 5 عبارات للتواصل اليومي

دليل خطوة بخطوة لممارسة المحادثة الإنجليزية

ما هي تقنية التظليل الصوتي؟