Taguette highlights: AI

Personalization & Adaptation

Depends on the kind of question I’m asking, like, if there are two reasons I
use chatgpt, sometimes it’s like I’m just having fun with it. I’m just asking
random questions or something that’s on my mind, which I want to know
and answer to. It can be some personal feeling or things like that. So in
moments like that, when I’m asking a question, I tend to talk to it like a
person, and I noticed that the answers are more personable, like it gives me
an answer, it feels more personal.

Document: interview_roger_otter_ai.txt Tags: AI as a Conversational

Agent, AI Personalization & Adaptation

I feel like, okay, based on recent interactions, I feel like it has a personality,
whichever you have.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Trust & AI’s Credibility

It adapts to your personality and gives you an answer in in a way that you
positively reinforce, for example, I sometimes, if I have a question in mind
and I want to use the mic function and give it a question, I say things like,
“Man, this was so bad”, or things like that. So it has picked up on that
personality. And now, when I say, for example, if I have a tough day, and “I
say I had a long day”, it’s like, “I hear you man,” or, “Oh, man, I feel bad
for you”. So I think it has picked up on my personality in a way.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Trust & AI’s Credibility
It does it. Sometimes, yes, it does it. I feel like it knows when it needs to
use an emoji also.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Trust & AI’s Credibility

Ah, no. I think I would still react. I would personally, personally, I think I
would communicate in the same way I do because um, because I have an
Alexa app, Amazon, Alexa and even to her, I just like I, I communicate in
the same way I communicate with chatGPT. And Alexa is less advanced.
It’s, you know, it doesn’t have a personality. So I find myself doing that.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

Yeah, because I think it keeps memory, and if I attacked it would remember

that.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

I think it does help me. It has helped me a lot. For example, writing emails,
I think is a social activity, or writing messages on WhatsApp, things like
that, where I feel like I need to be more diplomatic, and I don’t know how,
in those scenarios, I think chatGPT gives a very good answer.

Document: interview_roger_otter_ai.txt Tags: AI as a Tool, AI

Personalization & Adaptation

Do you ever feel like chatgpt understands you, or is it just an illusion of

understanding?
Speaker 2 8:50

I think it does understand you. Yes. I feel like yes in okay, if I can say this
in the interview, it feels like it understands your deepest fears and things
like that based on your all other conversations. So in that way you find you,
it feels like it understands you, maybe okay. It’s not a person to feel you,
but based on all the training and information it has, it knows that you have
these problems, and that is enough to know you, I feel. A little bit.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, AI & Emotional Perception

So then the next prompt, I say, “but this was also there in that situation”,
and then the answer changes. He’s like, “oh, this changes the dynamic”.
And then it’s a different answer. It’s more what is relevant.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

if chatGPT were to develop a unique personality just for you what traits
would you like it to have?

Speaker 2 11:36

Humor, sense of humor.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

es. Then it should know when to be serious, when not to.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation
Like I do really feel like chatgpt knows that, at least based on the
vocabulary I use in a question, it knows when it’s a serious question, when
it’s humorous.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Strategic Prompting & Role Assigning

In fact, I even have like, one separate chat, which I just did for humor
purposes and it instantly understood it was humorous, and it took a
completely different turn.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

And I even asked it like, “Okay, why did you take a humorous turn?” And it
said that “you use these words in your question because of that, I knew you
didn’t mean the question at all. You just did for fun”, so I was also having
fun with it.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Strategic Prompting & Role Assigning

I could give chatGPT to give me a nickname.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

For example, I use it for all kinds of stuff, solving problems, finding out
information, looking for research papers or telling it some problem, using it
as a personal therapist or whatever.

Document: interview_roger_otter_ai.txt Tags: AI as a Tool, AI

Personalization & Adaptation
But at the same time, I know some friends who use it purely for their work
or just purely to write code or things like that. So within between men and
women, I’m not sure. I think I’ve noticed some differences in men and
women like women, like several women have used it differently. So, yeah, I
think it should be like, I think everyone uses it the way they want or how
they need it.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Humanization of AI & Gender Perception

Politeness affects AI conversations? Um, I think it should affect to some

extent, like if you use emojis, then the ChatGPT will also use emojis. If you
are kind to it, then I think it uses the same kind of vocabulary to resonate
more with the user. So I think ChatGPT tries to give answers which the user
will most resonate to at that moment.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, Politeness & Directness in AI Interactions

Like real emotions. I think it’ll it shows real emotions by using phrases like,
“I hear you man”, or “I completely understand” or or emojis, or things like
that. It just makes you trust it more or like you feel like it understands you.

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation, AI & Emotional Perception, Trust & AI’s Credibility

“Can you answer this like I would?”,

Document: interview_roger_otter_ai.txt Tags: AI Personalization &

Adaptation

About the master’s thesis, how to build, and sometimes about the meal and
gym stuff, or groceries, listing and calculating all the macros, and maybe
translating German at that time. I send some answers to correct my German,
or if I have something in my head, I use it for language learning.

Document: interview_mel_turbo_scribe.txt Tags: AI as a Tool, AI

Personalization & Adaptation

I think it’s one of the weirdest things I can ever see, because it’s not real.
You can just train AI how to respond, so you’re just talking to yourself.

Document: interview_mel_turbo_scribe.txt Tags: AI Personalization &

Adaptation, Humanization of AI & Gender Perception

so you feel like it’s not genuine, authentic. Alright.

Mel: No, it cannot be, because you train how to respond.

Document: interview_mel_turbo_scribe.txt Tags: AI Personalization &

Adaptation, Trust & AI’s Credibility

Every day, specifically an hour ago, I wanted him to analyze my personal

constellation chart.

Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &

Adaptation

This is like associating everything with everything. It’s like personal with
them. Yeah.

Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &

Adaptation

You know, actually I asked one time like to generate my personality to

myself. And it was really harsh.
Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &
Adaptation

Yes, really harsh. And he called me like Carla, you are whirling this. What
is this called? Like this, you know, like one time he decided to quit and
smoking and you’re smoking behind your work.

Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &

Adaptation

I was like, it was using really bad language. So I was like, okay, that was
mirroring. Like how I was interacting with AI. It was incorrect.

Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &

Adaptation, AI & Emotional Perception

You are whirlwind of ambition drama, a touch of chaos wrapped up in a 44

kilogram package. You juggle sociolinguistics, teaching kids, working
customer service, writing academic papers, blah, blah, blah. And still find
the time to smoke behind.

Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &

Adaptation, AI & Emotional Perception

Interesting. Interesting. I’ve had then, you know, there was an answer. I will
not, I wouldn’t be able to find this conversation. He said that it seems that I
couldn’t satisfy you with my answers. And I was like, wow, I’m sorry. I
said, I think I had, I was sorry. I remember that I was like, oh, I’m sorry.

Document: interview_carla_turbo_scribe.txt Tags: AI as a Conversational

Agent, AI Personalization & Adaptation
If you could customize ChatGPT’s personality, what kind of character
would you give it?

Carla: Nerd

Document: interview_carla_turbo_scribe.txt Tags: AI Personalization &

Adaptation

I usually discuss. So most of my questions are related to ELT English

language teaching, and some of them are related to my spiritual side, and
most of them, I mean, I will say not some of them, and some of them are
related to my personal habits, routines, how to structure them, how to plan,
how to make weekly or daily plans to you know.

Document: interview_elvis_otter_ai.txt Tags: AI as a Tool, AI

Personalization & Adaptation, AI & Emotional Perception

Yes, I do, because I feel like when I assign a role to it, it starts acting in that
specific or particular way. So after that, if once I assign a role to it, after
that, it speaks exactly how that, how a person who has that role will speak,
you know what I mean?

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation, Strategic Prompting & Role Assigning

Okay, so in a way, you’re getting better response.

Elvis 3:51

Exactly, exactly.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation
And this was the question, do you think, given it the role influences the
response?

Elvis 4:02

Yes, I do. Yes, yes, I do.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation

I think I prefer giving an input first, explaining my thought process so that it

understands better what I have in mind. Because when I if I ask it like,
‘How can I be free?’ it can tell me millions ways of being free, but i want i i
want some ideas on how I, like me, can be free, because I have a different
perception of freedom, or whatever the subject is. You know, so it needs to
understand what I have in mind. It needs to understand my thought process
or my beliefs or my values so that it can give me a better response.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation, Strategic Prompting & Role Assigning

I actually try to observe that, but as far as I observed it, it’s all, it’s like the
same, it’s it just replies to the question or gives a specific answer to the I
mean, I don’t want to say question, but most of time I ask questions, so
yeah, but I don’t see any difference.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation

Exactly, exactly, exactly like it, its responses would improve. I think if I
provide emotional background.
Document: interview_elvis_otter_ai.txt Tags: AI Personalization &
Adaptation

When I share personal information, I think I yeah, yeah. I think the answer
I’m gonna give is going to be similar with one of the questions, but I don’t
remember the question, but I feel like if I share my personal background,
it’s gonna give me personalized response.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation

But if I just say, ‘What do you think about this dream without telling it,’ it’s
role. It’s just gonna, I feel like it’s just gonna give me a plain explanation.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation

Would you change anything about how AI responds to people responds to

you?

Elvis 20:52

No.

Document: interview_elvis_otter_ai.txt Tags: AI Personalization &

Adaptation

I believe it approaches everyone differently based on the data it has about

any of us. I mean, of course it’s gonna or we’re gonna say similar things, or
the conversations could be similar most of the time, but after a while, it gets
to know you, and it starts talking to you more personalized, I would say, so
it gets unique, you know.
Document: interview_elvis_otter_ai.txt Tags: AI Personalization &
Adaptation