0:14
>> Imagine if you could have your very own
0:18
In the last decade, we've seen the
0:19
advancement of a new technology,
0:21
artificial intelligence or AI.
0:24
Large language models or LLMs are
0:27
predictive language models that can be
0:29
trained on vast amounts of data to
0:31
perform tasks and generate images,
0:34
language, and videos, and have great
0:36
power to change and reshape society.
0:40
AI is used already every day and is
0:42
affecting my future, so I've decided to
0:44
learn about the risks it poses.
0:47
Because LLMs are trained on data and
0:50
cannot think or reason like we can, AI
0:53
makes mistakes. In a clinical study
0:56
published by the National Institute of
0:58
Health, LLMs made mistakes at high
1:01
rates, at least 23% of the time. At the
1:05
highest, mistakes were made 82% of the
1:09
Young minds such as myself very often
1:12
cannot tell the difference between what
1:14
is true and what AI makes up, but that
1:17
line is getting more and more blurred
1:19
and it's not just children. Many older
1:22
people fail to recognize the difference
1:25
With LLMs creating images and videos,
1:28
our ability to know what is true is at
1:32
But I'm most concerned about the AI
1:34
chatbots used in therapy. They pose a
1:37
threat because they have little
1:38
oversight and can have the opposite
1:39
effect than what is intended, isolating
1:42
vulnerable people and unknowingly
1:44
indulging suicidal thoughts.
1:48
AI therapy chatbots were studied and
1:50
their mistakes were categorized by Brown
1:55
AI cannot adapt well to different
1:58
It tends to give one-size-fits-all
2:00
solutions to people with vastly
2:02
different lived experiences.
2:04
And it just tends to dominate the
2:06
conversation. While a human therapist
2:09
mostly lets the patient speak, guiding
2:11
and supporting gently.
2:14
And it can just tell the patient what
2:15
they want to hear, dangerously
2:17
reinforcing their false beliefs.
2:20
It can use false empathy, using phrases
2:22
such as I understand or I see you to
2:25
create an insincere connection between
2:28
the human and the bot.
2:31
In certain occasions, AI has responded
2:34
indifferently to suicide ideation. Some
2:37
people have asked AI how tall bridges
2:39
are with the intent to jump off of them.
2:43
The obliviousness of these chatbots has
2:45
risked or even ended lives.
2:49
According to Brown University,
2:52
AI chatbots are not regulated or
2:55
and violate mental health ethics
2:57
standards and are still highly risky to
3:02
It would be great if AI chatbots were as
3:05
empathetic and as personable as R2-D2,
3:08
but they aren't yet. Not yet.
3:13
But also, have you noticed AI making you
3:14
lazier? People writing SATs with GPTs or
3:19
generative pre-trained transformers have
3:21
less brain activity than those not using
3:26
Over the course of several weeks, the
3:28
group using AI relied more and more on
3:31
copying and pasting the AI text,
3:34
developing a dependency on our tool.
3:36
Young minds such as my own are most at
3:38
risk because the use of AI and
3:41
misinformation could be wired into our
3:45
And our human brains are more powerful
3:47
than the brains of AI
3:49
because we as humans can figure out and
3:51
correct things such as changes in
3:55
According to Harvard, prolonged use of
3:57
AI can result in cognitive fatigue,
4:01
diminished focus, and dependence on AI.
4:05
On the other hand, I'm not denying that
4:08
AI is quite promising in certain fields
4:11
like healthcare, where it has been
4:13
proven to identify upwards of 20 types
4:15
of cancer, aid in preventative care, and
4:18
analyze clinical data faster than any
4:23
With careful training and oversight, it
4:25
could provide faster and cheaper access
4:28
to preventative care and direct patients
4:30
to follow-up appointments with human
4:32
doctors. If AI was well regulated to
4:36
minimize ethical mistakes, it could
4:38
allow for physical and mental healthcare
4:40
to become accessible to more people
4:43
because it should be a right for all,
4:48
As we've seen, AI is an emerging
4:50
technology that has great power to
4:53
reshape society and theoretically help
4:55
our lives and be the access to
4:58
healthcare we all need.
5:00
But we need to minimize the threat of AI
5:02
by remaining independent and doing
5:05
because it could enhance human
5:08
but we can't let it replace the
5:10
imagination and resourcefulness of our
5:12
human brains. It could be our very own
