Why Can’t We Use AI to Read Books?

00:00 In the age of AI,

00:02 many people oppose letting AI read for us.

00:07 And I wonder:

00:08 why do people say that?

00:12 Why has it become a slogan?

00:14 Personally, I think

00:16 AI reading for me is fine.

00:19 Sometimes, it's even essential.

00:25 If you listened closely,

00:28 you may have noticed—

00:29 or heard—a condition in what I said:

00:32 sometimes.

00:34 In other words,

00:35 whether AI should help us read

00:38 depends on the situation.

00:43 For example, sometimes

00:45 I need to achieve

00:47 a specific goal,

00:49 so I need to find

00:51 specific information in a book.

00:54 Could I have AI

00:57 do that for me?

01:00 Absolutely.

01:03 If I only need

01:05 certain information from a book,

01:07 and my goal is clear,

01:09 AI may get me there much faster.

01:13 Then our responsibility is

01:15 to judge whether the AI

01:18 can accurately

01:20 return the information I need,

01:23 and whether it might hallucinate.

01:27 There's another factor too:

01:34 the person using it.

01:37 Can they make good use of AI

01:40 to help them read?

01:43 Some people

01:44 have learned to think independently

01:47 and have the ability

01:50 to question things.

01:52 For them, using AI

01:53 to help interpret what they read

01:56 is completely possible, I think.

02:00 But for some people,

02:01 their abilities—

02:04 their ability to think,

02:06 or to question things—

02:08 may not yet be

02:11 always active in their minds.

02:14 They may also be

02:16 people who are still growing.

02:20 For people in that stage,

02:22 using AI to help them read

02:25 may unconsciously

02:28 weaken independent thinking

02:30 and weaken their ability to question.

02:35 That's one issue at the individual level.

02:40 So with all that said,

02:44 whoever we are,

02:48 when we use AI

02:51 to handle reading,

02:53 we should do it consciously.

02:59 And I don't think this applies only

03:03 to using AI for reading.

03:05 When we use AI tools—

03:08 any AI tool—

03:10 the same applies.

03:11 Whenever AI helps us

03:14 handle something,

03:16 we need to bring

03:18 awareness to the process.

03:20 I need to know

03:22 what goal the tool helped me achieve,

03:25 and what I gained

03:27 from using it.

03:29 And whatever that is,

03:31 we may need other abilities

03:34 to examine and interpret it—

03:37 judgment,

03:39 taste, and so on.

03:44 In an earlier episode, I talked about

03:48 outsourcing in the age of AI.

03:53 It was probably late last year.

04:00 I forget what I was thinking about then,

04:03 but it occurred to me that

04:04 there's a very popular

04:06 Chinese platform called Dedao.

04:10 As for exactly

04:12 what this platform does,

04:13 I won't try to summarize it,

04:15 because I don't know it very well

04:18 and have barely used the product.

04:21 So my interpretation here

04:26 may be wrong.

04:28 But my general impression is

04:31 that one section of its content

04:33 explains books.

04:35 It might invite

04:37 an expert to interpret a book,

04:41 or invite an expert

04:42 to create something like

04:46 a series on books or a course series.

04:51 Then readers,

04:54 or people pursuing "self-growth,"

04:57 consume these programs.

05:01 And that made me wonder:

05:04 these programs—

05:06 not only on this platform,

05:08 but similar programs elsewhere—

05:11 how are they different from AI book talks?

05:18 I think that's worth reflecting on.

05:24 In my own view,

05:26 if AI explains a book,

05:28 one possibility is

05:32 the issue we already encounter:

05:33 AI may hallucinate.

05:36 That's one point.

05:38 Another is whether its citations

05:41 and interpretations are accurate.

05:47 Following that line of thought,

05:50 the difference, to me,

05:53 between AI and a person explaining a book

05:56 is that AI has no perspective of its own.

06:00 Its content could be said

06:03 to come almost entirely from the book

06:05 and from the author.

06:07 A person, though,

06:09 has their own perspective,

06:10 their own life experience,

06:12 and their own values.

06:15 So the audience is effectively

06:18 seeing things through that person's world—

06:24 the book being discussed,

06:26 the course being taught,

06:29 or a person mentioned

06:31 in that course.

06:33 What I mean is,

06:35 I don't think this is

06:41 very different from outsourcing.

06:43 Why do we see things

06:44 in our society

06:46 like KOLs and "internet mouthpieces,"

06:50 or people asking others for

06:52 book lists,

06:53 movie lists,

06:54 things like that?

06:55 At heart, they're closely connected

06:57 to outsourcing in the age of AI.

07:05 Personally,

07:08 if many audience members

07:14 want to grow

07:16 and rely on platforms

07:19 and programs like these to do so,

07:23 they are essentially outsourcing

07:26 some of their own abilities:

07:28 like

07:29 their ability to filter,

07:31 to interpret,

07:33 to think,

07:35 taste,

07:36 their judgment, and many others

07:38 I haven't even mentioned.

07:41 If that's the case,

07:44 whether in today's

07:47 AI-powered world

07:48 or in the era before AI,

07:50 this kind of outsourcing

07:53 has never been rare.

07:57 AI has simply made

07:59 the phenomenon

08:02 easier for us to notice.

08:05 If that's really true,

08:08 some people loudly insist that

08:11 our taste

08:12 and our judgment

08:13 must never be outsourced.

08:15 But what use does that slogan

08:17 still have

08:20 in our present era?

08:22 I don't know.