本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
创意社区终于开始正视AI在创作过程中的使用问题，而这全是因为Hank Green在最近一个视频里说的四个字。
The creative community is finally addressing AI usage in the creative process, and it's all because of four words Hank Green said in a recent video.
我很感激这种反对的声音。
And I appreciate the pushback.
这四个字在Hank的社区乃至整个互联网上引发了众怒，也导致Hank暂停了他所有频道的更新。
These four words sparked outrage amongst Hank's community as well as just across the entire internet and led to Hank pausing uploads on all of his channels.
关于AI在创作过程中的使用，这个话题我们在办公室里和团队讨论过很多次。
Now this conversation around AI usage in the creative process is one that we've had quite a bit in our office with our team.
但自从Hank Green那件事之后，过去一周这个话题的讨论热度明显上升了。
But that conversation has picked up a lot over the past week since this moment with Hank Green.
所以这期节目我们请来了不同背景的嘉宾，一起聊聊这条界限。
So we wanted to bring a bunch of different voices into the room for this episode where we talk about the line.
创作者应该如何在创作过程中使用AI？
How should creators be using AI in their process?
让我们开始吧。
Let's get into it.
好的。
Alright.
所以这个Hank Green的事情已经完全炸开了。
So this Hank Green situation has completely blown up.
我真的很惊讶有这么多媒体都报道了这件事。
I'm actually shocked at how many publications have written about it.
太疯狂了。
It's crazy.
我的意思是，视频发布大约一周后，The Verge十五分钟前刚发了一篇文章。
I mean, we're about a week after the video was posted and The Verge just post an article about it fifteen minutes ago.
这个话题还在持续讨论中。
It continues to be a topic of discussion.
那我们来解释一下发生了什么。
So let's explain what happened here.
好的。
Okay.
简单说一下，如果你不了解汉克·格林和他的故事，就快速介绍一下，他和他的兄弟约翰，他们是作家、作者、学者，而且他们在YouTube上已经二十年了。
Briefly, if you don't know Hank Green and his story, just quickly, he and his brother John, they are authors, writers, academics, and they have been on YouTube for twenty years.
他们是传奇人物。
They are legends.
他们有一个频道叫Vlogbrothers，在那里他们互相发送视频。
They have a channel called Vlogbrothers, where they send videos back and forth to each other.
现在，基于这个成功，他们创办了一家公司叫Complexly，旗下有Crash Course、SciShow。
Now, off the back of that success, they've started a company called Complexly, which has Crash Course, SciShow.
这是一系列教育视频，在全国的高中里播放。
It's a series of educational videos, which are shown in high schools throughout the country.
他们主要被大家熟知为教育者，基本上就是互联网上的老师。
They are known largely as educators, like the internet's teacher, basically.
现在，这一切开始的那个节目叫《Ask Hank Anything》，在Complexly的YouTube频道上，YouTubers会来问汉克问题。
Now, the show where all of this started is called Ask Hank Anything on the Complexly YouTube channel, where YouTubers come in and they ask Hank questions.
我们被邀请上过这个节目两次。
We've been asked to be on the show twice.
我不太清楚。
I don't yeah.
我们没成功吗？还是怎么回事？
We it didn't work out or what?
是我的错吗？
Was it my fault?
是的。
Yes.
好的。
Okay.
抱歉，汉克。
Sorry, Hank.
所以在这一集里，一位叫Supey Time的创作者在向汉克提问。
So in this episode, a creator named Supey Time is asking Hank questions.
这是一个长格式的播客。
It's a long form podcast.
这个节目很有教育意义。
It's very educational.
然后他们开始深入探讨词语的含义，以及词语是如何得名的。
And they start getting into the meaning of words, how words get their names.
接着切换到一段汉克教你知识的环节。
It throws to a segment where Hank is teaching you something.
他在解释两个词是怎么得名的。
He's explaining how two words got their names.
而正是这段剪辑引发了所有讨论。
And this is the clip that sparked everything.
这里面有个关键点，而且非常有意思。
There's a thing in there that matters, and that's very cool.
更好的说法是——我很感谢你的指正——词是人类发明的，而人类的发明是由身体和大脑利用它们所拥有的系统共同创造的，包括
A better way to say it, and I appreciate the pushback, is a word is a human invention, and human inventions are made by bodies and brains working together with the systems that they have, including
所以关键的时刻就是他说“我感谢你的指正”那一段。
So the moment in question is him saying, I appreciate the pushback.
更好的说法是，我感谢你的反驳。
A better way to say it, and I appreciate the pushback.
对。
Yeah.
所以现在重要的背景是，之前他们在争论，她在反驳他。
So now important context is like prior, they are debating, and she is pushing back on him.
但对当时情况的理解是，Hank在读一个ChatGPT生成的脚本，因为“我感谢你的反驳”听起来不自然，像是AI写的东西。
But the read of what happened there is that Hank was reading a ChatchyBT created script because I appreciate the pushback sounded unnatural and sounded like something that was written by AI.
这很容易
It would
让人以为Hank在和ChatGPT对话，要求一个更简单的解释版本。
be easy to assume that Hank was having a conversation with ChatGPT, asked for a simpler version of this explanation.
嗯。
Mhmm.
然后ChatGPT回复说，另一种说法是。
And ChatGPT responded saying, another way to say it.
我感谢你的反驳。
I appreciate the pushback.
嗯。
Mhmm.
然后给出了更简单的解释。
And then gave the simpler explanation.
所以这就是大家的猜测。
So that's the assumption.
现在这个视频火了。
And now this clip goes viral.
他的粉丝开始感到不满，觉得汉克在给他们读的内容中使用了AI。
His community starts to really feel upset that they feel like Hank's using AI in something that he's reading to them.
然后汉克发了一条推文，否认那句话是AI写的，说那是拍摄时的即兴发挥，但承认他在高压下制作视频，因为视频量大，确实用ChatGPT做了脚本研究。
And then Hank puts out a tweet where he denies that that specific line was written by AI, saying it was an ad lib during the filming, but acknowledges that he's made videos under heavy pressure because he's making a high volume of videos and has used ChatGPT for research on the script.
他随后删除了那条推文，又发了一篇更长的说明。
He then deleted that tweet and then posted a much longer note
在他的subreddit上。
on his subreddit.
这个东西像是一大页，感觉有点像道歉和承认。
This thing is like a big page, almost feels like somewhat of an apology and an admission.
这确实是Hank对自己使用LLM程度的一次深刻反思。
It is very much a lot of introspection from Hank about how far his use of LLMs has gone.
他确实为可能因此与观众产生隔阂而道歉，并表示他会退一步。
He does apologize for how it's potentially gotten in between him and his audience, and he says that he's going to be taking a step back.
他打算在Hank的频道和其他几个频道减少视频制作，直到他完全理解自己的视频制作流程。
He's going to be making less videos on Hank's channel, a couple of his other channels, until he starts to understand fully his process for making videos.
现在那里有一句话，我觉得每个人，从我在YouTube和主播那里看到的反应来看，都对此有反应。
Now there's a line in there that I think everyone from what I've seen all over YouTube and streamers have reacted to this.
大家都在引用这句话，他谈到自己与LLM互动时获得的多巴胺。
Everybody is quoting this one line where he talks about the dopamine he's getting from interacting with LLMs.
这开始涉及到他承认自己某种成瘾的状态，人们对此有所理解。
And that starts to play into this world of him kind of admitting a version of an addiction that people are understanding.
现在，这件事，汉克·格林以任何形式使用AI，已经引发了巨大的愤怒。
Now, this thing, Hank Green using AI in any capacity has created so much anger.
我截了一张评论的图，现在他在YouTube上所有视频的评论区里，都是对他感到愤怒的人。
I screenshotted a comment from The comments on all of his videos on YouTube now are just people angry at Hank.
对吧？
Right?
所以，比如，这是他今天早上发的最后一个YouTube视频下的评论，而这个视频跟这件事完全无关。
So, like, this is a comment from this morning on his last YouTube video that has nothing to do with this.
搞什么鬼，老兄？
What the hell, man?
你是我为数不多信任的、保持真实的创作者之一，结果你现在却在这里给我们喂AI垃圾，还把它当成你自己的话来发。
You are one of the only creators I trusted to be authentic, And you're out here feeding us AI slop, passing it off as your own words.
如果你时间紧张，你可以雇个人帮你写稿子，或者帮你做研究，或者跟你一起头脑风暴。
If you were under a time crunch, you could have hired someone to help you with your scripts or to do your research or to bounce ideas off.
无论如何，使用生成式AI都是没有借口的。
There's no excuse for using Gen AI in any capacity.
这不是不可避免的。
It's not inevitable.
这不是未来。
It's not the future.
这不是工具。
It's not a tool.
这只是给那些不会独立思考的人准备的垃圾内容。
It's just slop for people who can't think for themselves.
而且最糟糕的是，我再也不能相信你会独立思考了。
And it majorly sucks that I can't trust you to think for yourself anymore.
这感觉太过了。
That feels really far.
对。
Yeah.
我看Hank的视频，即使是那段剪辑，我也没觉得这跟我认识和喜欢的那个Hank Green有什么不同。
When I watch Hank, even that clip, I do not feel like it is anything different from the Hank Green that I know and love.
也许在私下里，即使他像他说的那样用ChatGPT来制作那个视频，对我来说，他还是我认识和喜爱的那个伟大的Hank Green。
Now maybe behind closed doors, even if he is using ChatGPT as he says to develop that video, to me, it feels it's the same great Hank Green that I know and love.
所以，是的，我们对AI的看法不同，但知道他用AI并不会影响我对他的看法。
So, yeah, I he different views on AI, but, like, it does not knowing that he uses it does not tarnish my opinion
不会。
No.
不会影响。
Of it.
对。
Yeah.
这个反应让我很惊讶。
This this reaction has surprised me.
我的意思是，我知道大家对AI的普遍看法，现在我想深入聊聊这个。
I mean, I'm aware of the general sentiment towards AI, and I wanna get into this now.
而且我们房间里还有其他人的声音。
And we have other voices in the room.
我们的故事制作人康纳也加入了我们。
We're joined by our story producer, Connor.
还有来自出版媒体的作家席德·科恩，以及新加入团队担任创意策略师的基什。
We're also joined by writer from the published press, Sid Cohen, as well as Kish, who's new to the team as a creative strategist.
你们俩，那边那两位，你们多大了？
The, you two over there, what is your guys' state your age.
我23岁。
I am 23.
你是20岁，基什是……我20岁。
You're 20 Kish is I'm twenty
26岁，而且不再是房间里最年轻的人了。
twenty six and no longer the youngest person in the room anymore.
对。
Yeah.
所以我觉得聊聊这个很有意思。康纳，你多大？
So I think it's interesting to talk And Connor, your age?
30岁。
30.
30岁。
30.
有意思的是，我和Colin都是37岁。
So it's interesting because Colin and I are both 37.
所以，我们和这个话题的不同关系，我觉得真的很有意思。
So to talk about our different relationships with this subject matter is, I think, really interesting.
对于生成式AI，有一种愤怒，这种愤怒我其实一直不太理解，尤其是来自Z世代的。
There is an anger towards generative AI that is actually, has been hard for me to understand from Gen Z.
Syd，你能解释一下我刚才读的那条评论吗？这种愤怒是从哪里来的？
Syd, can you explain, like the comment I just read, where is this anger coming from?
有一件重要的事情我们还没提到，就是他澄清说，Hank是用AI来汇总研究论文，然后他自己去读这些论文。
Something important that we have not touched on yet is his clarification that he that Hank was using AI as an aggregator for research papers that he then read.
所以并不是用AI来写他的稿子。
So it was not that he used it to write his script.
他是把它当作一个工具，用来获取那些他可能通过谷歌学术或者去图书馆都找不到的研究资料。
He used it as a tool to get to research that he might not have gotten to otherwise through like Google Scholar or going to the library.
这个工具帮他找到了资料，然后他自己读了那些内容，并且他公开说了这件事。
This was a tool used to get to a product that he then read himself and he came out and said that.
所以我觉得这个澄清很重要，因为这影响了我对现在看到的那些反对声音的看法和感受。
So I think that clarification is important for the context of my opinion and how I feel about the backlash that I've been seeing.
但从Z世代的角度来看，我理解他们为什么生气。
But from a Gen Z perspective, I think why they're mad.
我从小就是看汉克的视频长大的。
Like I grew up watching Hank.
他帮我通过了AP心理学考试。
He got me through AP Psychology.
我15岁的时候就用他的视频当学习工具，那时候AI都还不存在呢。
Like I was 15 using his videos as a tool when AI didn't even exist yet.
所以现在看到大家这么不满，我完全理解。
So now to see people upset, I understand.
我理解这种愤怒。
I understand the outrage.
我觉得这是因为在讨论中，大家对什么算AI没有明确的区分。
And I think it's because there's not a clear distinction in the discourse on what constitutes AI.
我觉得AI的定义太模糊了。
I think the definition of AI is so nebulous.
比如，每个平台都有AI工具。
Like, every single platform has an AI tool.
用生成式AI写剧本和生成视频，跟用它来汇总数据然后自己去做研究，是完全不同的两回事。
And the the difference between Gen AI for writing a script and generating a video is completely different than using it as a way to aggregate data, which you can then use to do your own research.
但人们生气是因为他们的信任被所谓的“背叛”了。
But people are mad because their trust has been quote unquote betrayed.
但再深入一点，AI到底代表什么？
But but go deeper on, like, what does AI represent?
我觉得有两方面。
I think there's twofold.
对吧？
Right?
一边是汉克·格林，一边是AI。
There's Hank Green and then there's AI.
汉克·格林是教育者。
Hank Green is the educator.
他经常处于权威位置，人们会参考他的观点。
He's this guy who's often in the position of authority that people look to for their opinions.
说实话，很多人都是看汉克怎么说，然后形成自己的看法。
A lot of people, if we're being honest, look to Hank to assign them opinion on many topics.
现在他们被告知，很多话题上给他们的观点其实来自AI。
And now they're being told that the opinion that was being assigned to them on so many topics came from AI.
然后还有Z世代对AI的看法。
And then there's how Gen Z views AI.
想想看，现在的Z世代，如果你22岁，刚上完大学，花了那么多钱，现在可能还没工作。
Think, Gen Z right now, if you're 22, you just went to college, you spent all this money, and now you're probably unemployed.
你很可能失业了，或者没法搬出父母家，心里很窝火。
You probably are unemployed or you can't move out of your parents' house and you're angry.
然后你上网一看，大家都在说AI市场巨大，股市创历史新高。
Then you go online and they're like, AI has this big market, stock market, all time high.
AI在创造就业机会。
AI is bringing in jobs.
或者说AI会抢走你的工作，然后各种说法来回拉扯。
Or is AI going to take your job and then it's going back and forth.
然后你就会想，汉克·格林，这个互联网的叔叔，明明请得起研究员，却用AI。
Then you're like, Hank Green, this guy, the Internet's uncle is now using AI when he could clearly afford a researcher.
对吧？
Right?
就像如果我们说，什么
Like if we're What's
你对研究员会做什么有什么预期？
your expectation on what the researcher would do?
对。
Yeah.
你明白我的意思吗？
Do you know what I'm saying?
我在这里反驳是因为，即使是在用AI做研究的前提下，从“我用ChatGPT或某个大语言模型来整理思路、更快地获取信息，比我去图书馆还快”到“我再也不能信任你了”之间，真的有很大的差距。
And I'm pushing back here because I'm like, even in the premise of using it for research, there's a really big gap between I used ChatGPT or some LLM to organize my thoughts and provide me with information faster than I would've got if I walked to a library, and I can't trust you anymore.
对我来说，这是一个巨大的差距，我想更深入地理解，从“我用AI做研究”到“你说什么我都不能信”之间到底发生了什么。
That to me is like a massive gap that I'm trying to understand deeper on, get me from, I used AI to research this, to, I can't trust you with anything you say.
针对Keisha的观点，我觉得这里的核心是思考被外包了。
To Keisha's point, I think the idea here is that thought is being outsourced.
我自己也有这种感觉。
And I see that for myself.
比如，我是你通讯稿的撰稿人，我们的通讯稿。
Like, I'm I'm a writer for your newsletter, for our newsletter.
嗯。
Mhmm.
我在写作过程中不用AI大模型，因为我觉得我的工作就是把创作者经济里那些想法简洁地整合起来，给读者提供一个观点。
And I don't use AI LLMs in my process because I feel that it is my job to succinctly synthesize the ideas that are floating around in the creator economy and provide a, like, point of view to our readers.
对。
Mhmm.
所以我不想让这个过程的任何部分被外包出去。
So I don't want any part of that process outsourced.
你觉得他用工具找到了一个别处找不到的资料，这就算外包思考吗？
Do I think that by him using a tool to find a source that he might not have been able to find elsewhere is necessarily outsourcing thought?
我不这么认为。
I don't think so.
但我也理解为什么有人会觉得，在过程中任何环节用了大模型，就是把你的工作外包给机器而不是人。
But I also understand why someone would think that an inclusion of an LLM at any point in the process could be outsourcing your work to a machine instead of a human.
我觉得这就是大家生气的地方。
And I think that's where people are getting mad.
但你觉得所有这些批评都合理吗？
But do I think that all the hate is warranted?
当然不是。
Absolutely not.
我们在已发布的媒体上做过一个投票，问大家创作者在什么情况下应该披露他们在创作过程中使用了AI。
We have we ran a poll in the published press asking people when it's okay to when when creators should disclose that they've used AI in their process.
康纳有完整的引述，但你能读一下关于道德标榜的那部分吗？
And Connor has the full quote, but could you read the one about virtue signaling?
好的。
Yeah.
好的。
Yeah.
所以投票的一部分，我觉得选项是：创作者应该在什么时候披露他们在内容工作流中使用了AI——总是在面向受众的用例中，比如图片、视频、脚本写作，或者从不，或者写一个选项。
So part of the poll, I think the the options were to answer when when should creators disclose AI in their content workflow always in audience facing use cases, which would be like images, video, script writing, never or write an option.
这个人只投了面向受众的用例，然后回复说反AI是最新的道德标榜，我这么说是因为我支持反AI。
And this person had had voted only in audience facing use cases, And the response was anti AI is the newest virtue signal, and I say this as someone who supports being anti AI.
整个问题在各方面都是一团伦理乱麻。
The whole problem is an ethical mess on all sides.
嗯。
Mhmm.
对。
Yeah.
所以，所以，所以，这就是那个人的观点，认为环境问题、失业问题、小镇上的数据中心，基本上就是对你人品的评判。
So so so that's someone who's taking the the POV there that the environmental concerns, the job loss, the data centers in small towns, that it's a judge of your character, basically.
对。
Yeah.
就是说，这个东西本身因为它的存在方式就是不道德的，而不是说使用它不道德。
That the thing itself is immoral because of how it exists, not the use of it.
所以我觉得，之所以要讨论这个话题，并且围绕创作者来讨论，是因为在这个对话中确实需要一些细微的差别。
So I think it's important, the reason to have this discussion and have this discussion around creators is because there's a necessity here for some levels of nuance in this conversation.
现在有很多非黑即白的观点。
There's a lot of black and white.
有很多人，比如Hank，在创作过程中某个环节用了LLM。
There's a lot of Hank used an LLM at some point in his process.
汉克现在对我来说就不算数了。
Hank now not valid to me.
我觉得这真是我见过的一种极端的非黑即白。
Like, that is a incredible version of black and white that I'm seeing.
而且特别有意思的是，我个人不认为这种情况会发生在每个创作者身上。
And super interesting because I don't personally believe that this situation would happen to every creator.
我觉得不会发生在你们身上。
I don't think it would happen to you guys.
对。
Yeah.
我觉得我们一直对外公开我们与技术的关系，但我也认真思考过这个问题，也想过我们之间的年龄差异。
I think we've been outward facing about our relationship to technology, but I have also thought a lot about this and thought about our difference in ages.
我和科林都是37岁，我们大学毕业那会儿，我们的同龄人正在硅谷开发应用，他们采用技术的方式改变了他们的生活，在我们眼里他们就是明星。
So Colin and I, being 37, when we came out of college, what was happening was our peers were creating apps in Silicon Valley and adopting technology in a way that was changing their lives, and they were rock stars to us.
你可能很难相信这一点。
This may be hard for you to believe.
对我们来说，马克·扎克伯格就是摇滚明星一样的存在。
Mark Zuckerberg was a rock star to us.
他们的应用改变了我们的生活。
And their apps were changing our lives.
我们在学校里认识新朋友。
We were meeting people at school.
我们在约会。
We were dating.
我们就是那些人，做Tinder的，做Instagram的，还有那些，那是一场运动，Evan Spiegel做了Snapchat，年轻人用技术改变世界。
We were people, the guys who made Tinder, the guys who made Instagram, the guys, there was a movement, Evan Spiegel, made Snapchat of young people utilizing technology to change the world.
在我们21岁的时候，那真是让人无比兴奋。
And that was extraordinarily exciting when we were 21.
我现在依然有这种感觉。
I still have that.
我们这一代人的信念就是，尽早拥抱技术可以解锁并改变你的世界，甚至改变整个世界。
The thing of our generation is being an early adopter of technology can unlock and change your world and change the world.
所以我和科技的关系就是，每当有新科技出现，我都想成为最早使用它的人。
So my relationship with technology is when a new technology comes my way, I want to be an early adopter of it.
而且我想尽可能多地了解它，因为我从小就看到这些。
And I want to learn as much of it as possible because I witnessed that growing up.
不过，我并不是在评判这是好是坏。
Now, I'm not assigning good or bad to that.
我只是在分享我自己的经历。
I'm just assigning my own experience.
我想和我同龄或者更年长的人对科技的看法应该也差不多。
And I would assume people in my age demographic and older feel the same way about technology.
我觉得像你这种年龄段的人，其实对科技挺信任的。
I think people in your age demographic though, have a lot of trust.
我们这一代人就不一样了，我们不相信政府。
Our generation does not We don't trust the government.
我们也不相信我们的——我差点说我们不相信老板，但是……
We don't trust our I was about to say we don't trust our bosses, but
我不信任
I don't trust
一个。
one.
我们就是不信任这些人。没错，
We do not trust these guys And right
虽然我理解你的反驳，但我觉得公平地说，Hank不只是说他做了研究。
while I appreciate the pushback, I think think what is fair to say though, is that Hank did not just say he researched.
他还继续说，他沉迷于LL。
This man went on and said, am addicted to LL.
对。
Yeah.
所以这就是我们讨论的第二部分。
So that's part two of this discussion.
对。
Yeah.
这不像我们那位，他只是说，对，我想在研究员上省点钱。
And it's not like our guy where he just like, he's like, yeah, I wanted to save money on a researcher.
这就像，首先，如果你知道自己运营的规模跟不上，那就招人或者想个流程出来。
It's like, for one, if you know you're operating at scale you can't keep up with, then bring on staff or figure out some process.
还有，就像我说的，对，你也许可以说，把这定义为他的性格是不对的。
Also, like I was saying, it's like, yeah, you probably could say it's invalid to say that this is a definition of his character.
但说你对AI上瘾，这确实是对性格的定义。
To say you are addicted to AI is a definition of character.
澄清一下，他没说上瘾。
To clarify, he didn't say addicted.
是我们理解成上瘾了。
We're reading addicted.
对吧？
Right?
他，他没说上瘾。
He he didn't say addicted.
他说的是Verizon多巴胺。
He said Verizon dopamine.
对。
Yeah.
你能说一遍吗？
Can you say it?
可以。
Yeah.
他说的是，但主要是，我需要接受一个事实，那就是我一直在和LLM互动中得到的多巴胺水平，越来越多，越来越多，对我来说不健康，对世界也不好。
He said, but mostly I need to come to terms with the fact that the level of dopamine I've been getting from interacting with LLMs with doing more and more and more and more is not healthy for me or good for the world.
好的。
Okay.
对。
Yeah.
那有点不一样。
That's a bit different.
对。
Yeah.
你看，我觉得网上经常会有这种跳跃。
See, I think there's all there's often a jump in in the Internet.
对吧？
Right?
因为，当然，人们只是随便看一眼，所以他们一看到那个就马上说，Hank Green 对 AI 上瘾了。
Because, like, of course, like, people people just glance at things, and so they immediately glance at that and go, Hank Green's addicted to AI.
他什么都用 AI。
He's he uses AI in everything.
他说的话没有一句不是 AI 生成的。
He doesn't say a word that's not AI.
我觉得这种跳跃挺夸张的。
Like, there's a jump that I think is is dramatic.
而且，对。
And yeah.
请讲，先生。
Go ahead, sir.
我正想说，回到你刚才提到的关于技术代沟的观点，我觉得当时你们之所以那么兴奋，是因为那时候的技术感觉真的很无害。
I was just gonna say kinda going back to your point on, like, the generational gap with technology, I think that the reason why it was so exciting for you guys is because the technology at the time felt really benign.
就是，感觉不会威胁到生命。
Like, it didn't feel life threatening.
那时候还没有算法。
There were not algorithms yet.
就像，我要发这张照片，然后按时间顺序看到所有东西，感觉就像一种方式。
It was like, I'm gonna post this picture, and I'm gonna see everything in chronological order, and it feels like a way
相对而言吧。
Relatively.
我是说，YouTube 当时就有算法，它为我打开了世界的大门。
I mean, YouTube was an algorithm and it I unlocked the world for
我觉得也许回到 Sid 的观点，这些科技公司的规模已经……
think maybe to Sid's point, like the scope of these tech companies has
对。
Yeah.
变了。
Changed.
变化非常大。
Has changed dramatically
我们18岁时用的Facebook和现在的Meta相比，变化真是天翻地覆。
in- What Facebook was when we were 18 to what Meta is today is dramatic.
那时候它就是个好玩的地方，可以跟朋友发照片，还能认识其他学校的人。
Like it was a cool place to post photos with your friends and to like meet people from other high schools at the time.
而且还能……
And to
你说得对，那时候马克·扎克伯格就像个摇滚明星，而你们这一代人大概从小就是看他在国会作证长大的。
your point, like, Mark Zuckerberg was a rock star, and I think your generation probably grew up seeing him testifying in front of congress.
是的。
Yes.
这差别真的很大。
It's like very different.
对。
Yeah.
我是看着马克·扎克伯格的蜥蜴表情包长大的。
I grew up with Mark Zuckerberg lizard memes.
对。
Yeah.
没错。
Right.
没错。
Right.
然后，对。
And Yeah.
但我觉得，现在，像我这一代人看科技公司的时候，看到的全是人们家里连水都开不了，因为水是易燃的，就因为旁边在建数据中心。
But I think and I think now, like, when I view a tech company, when people in my generation view a tech company, it is all I'm seeing is people can't turn on the water in their homes because it's flammable because there's a data center being built next door.
有些人的照片在未经同意的情况下就被使用了。
People are having their images used without their consent.
他们呢，我想就像你说的，因为我们之前聊过这个话题，我敢肯定他们和我们一样都签了条款协议，但他们并不希望事情变成这样。
They are well, I guess to your point, because we've had a conversation about this, I'm sure they signed the terms of conditions like the rest of us, but not wanting that to be the outcome.
对。
Right.
我觉得这些科技公司，也许只是我年纪大了，但我在和它们打交道时，感觉越来越有一种不祥的预感。
I think that these tech companies, maybe it's just me getting older, but it's become more of a sinister feeling that I get when I'm interacting with them.
这也加深了我对人工智能、大语言模型和社交媒体平台的不信任，尽管我还在继续使用它们。
And that's fueled my mistrust or distrust towards AI, LLMs, social media platforms, even though I do continue to use them.
我说的它们，指的是社交媒体平台。
And by them, mean the social media platforms.
对。
Yeah.
我听到的，不知道对不对，就是这种情况，或者任何公开在创作过程中使用AI的创作者，都和黑暗面有某种关联。
There's an alignment with like, from from what I'm hearing, correct me if I'm wrong, but this situation or any creator who's, like, outwardly using AI in their process, there's an alignment with the dark side.
因为你的假设是这就是黑暗面。
Because your assumption is this is the dark side.
对吧？
Right?
调查里有一个很有意思的评论，有人说创作者应该始终坦诚地说明是否使用了AI。
There's a there's a really interesting comment from the survey, and this is somebody that said creators should always be upfront about AI use.
而且因为所有AI模型的创作、训练和使用都涉及伦理和道德问题，考虑到AI依赖知识产权盗窃以及数据中心对环境的局部负面影响，创作者坦诚地说明AI的使用程度才是公平的，这样我才能决定以什么程度参与你的内容。
And because there are ethical and moral issues baked into creation, training, and use of all AI models, given AI's reliance on intellectual property theft and localized negative environmental effects of the data centers, it's only fair that creators are honest about how much AI is present so I can make my decision on what level of engagement I have with your content.
完全同意。
100%.
我们来设想一下，如果Hank写了很多想法，然后让一个LLM帮他整理这些想法，会怎么样。
Let's let's play out the world where Hank writes a bunch of thoughts and ask his asks an LLM to organize those thoughts for him.
对。
Yeah.
那这算什么呢？
Where does that lie?
我觉得这算是踩线了，可能算黄色区域吧，但你应该还好。
I mean, that that is like on the line of like it probably be in the yellow territory, but you probably like, you're good.
但如果这家伙用得太多，已经渗透到他的用词里了，或者他常说“我感谢你的反驳”。
But I mean, if this dude is using it so much that it has baked into his words or he used to say, I appreciate the pushback.
我觉得这个说法，没多少人会真的说出口。
I mean, that's a phrase that I don't think many people ever say.
比如
Like
我们现在是认真的吗？有人连“我感谢你的反驳”都不能说？
What do you are we being for real right now that somebody like, you can't say, I appreciate the pushback?
兄弟，AI每次都用这个说法。
Bro, like AI uses that phrase every single
这其实又回到了我们一直在聊的关于AI的一个点：现在AI更像是一种感觉，而不是一个明确的东西。
this really goes back to something we've been talking about a lot about AI at this point is way more of a feeling than, like, a definitive thing.
就像，即使我们在争论，如果你口头说“不是这个，是那个”，现在这已经太AI化了，尽管这是一种非常常见的论证方式，但现在它反而成了你用了AI的信号，哪怕你其实没用。
Where it's like, even if we're making an argument and if you verbally say, like, it's not this, it's that, it's like so AI focused now that despite being, like, an incredibly common way to make an argument, it's now just a signal that you used AI even if you didn't.
这完全可以成为你的日常用语。
That can just be, like, in your vernacular.
就像你不能再用破折号了一样。
It's like you can't use an em dash anymore.
完全同意。
Totally.
对。
Yeah.
我在这方面挺差的
I'm pretty bad
确实。
at that.
我就要用那些破折号。
I'm gonna use those em dashes.
这是个好习惯。
It's a good grim.
但我觉得这挺有意思的，就是说，好吧，假设在一个世界里，汉克没有这么做，但假设他完全按稿子来拍那个视频。
But I think this is an interesting point that like, okay, let's live in a world where and Hank did not do this, but let's live in a world where he scripted that video completely.
但那段关于反对意见的话没被剪进去，也没人注意到。
But that line about pushback wasn't included and no one noticed.
我觉得现在很多创作者在规模化的时候都会遇到这种情况。
That's happening in my opinion with a lot of creators at scale.
不是你的，我不会说这是我们的看法，甚至我们都已经看到了。
Not in your I wouldn't say in our opinion even like we've we've seen it.
在我们所知道的范围内。
In what we know.
对。
Yeah.
我们知道。
We know.
对吧？
Right?
大多数创作者都会以某种方式在工作中使用AI。
Most creators use AI in their workflow in some way and another.
就像我刚才说的，我在我的流程中不用大语言模型。
We like, I just said, I don't use LLMs in my process.
但我确实用转录服务，因为我不想坐在那儿听音频。
I do use a transcription service because I'm not going to sit there and listen to the audio.
对着镜子打一个小时电话。
Call in his mirror for an hour.
不。
No.
我永远不会听你们的。
I would never listen to you guys.
我觉得有意思的是
I think what's interesting is
如果用例足够好，或者梗足够好笑，没人会在意。
if the use case is good enough or if the meme is funny enough, no one gives a shit.
对。
Yes.
我们这周一直在聊这个话题，所以想把它带到这儿来。
We've been talking about this all week, which is why we wanted to bring this here.
但我记下来了，你告诉我这是不是其中的一部分。
But I wrote down Just tell me if this is part of it.
我能理解，AI在某种程度上确实对工作构成了威胁。
I can understand this, that AI represents this threat to work in a way.
对吧？
Right?
但它绝对对年轻人的工作构成了威胁。
But it definitely represents a threat to work for young people.
是的。
Yep.
然后是对诚实的威胁，我觉得这个很有意思，就是诚实受到的威胁。
And then a threat to honesty, which I find to be a really interesting one, that the threat of honesty.
我之所以写下这个，是因为创作者经济的承诺是，你看到的是一个人真实的一面。
And the reason I wrote that down was because the promise of the creator economy is that you are experiencing an honest version of someone.
对吧？
Right?
是的。
Yeah.
就像看汉克的时候，至少对我来说，这并不会削弱汉克·格林作为朋友、创作者和同代人在我心中的真实感。
Is that and like with Hank, when you're watching him, at least for me, like and I this to me does not take away that reality of Hank Green to me as a friend, as a creator, as a contemporary.
我觉得汉克是YouTube上乃至我们这一代最伟大的思想家、作家和表达者之一。
Hank, I think, is one of the greatest thinkers, writers, and articulators on YouTube and of our generation.
我觉得现在发生的是，对某些人来说，我刚才说的话已经不可信了，因为他做的事情缺乏真实性，那不是他脑子里想出来的，而是和别的东西合作的结果。
And I think what's happening is that to some people, what I just said has been discredited because there's the lack of honesty to what he did, that it's not thoughts out of his brain, it's thoughts in collaboration with something else.
我想分享一个相关的东西，是《纽约时报》的一篇文章。
One thing I do wanna share that is relevant is an article from the New York Times.
标题是《如果你超过40岁，你已经准备好使用AI了》，作者是一位设计师兼教授，名叫黛比·米尔曼。
It's titled, If You're Over 40, You're Ready to Use AI, written by a designer and a professor named Debbie Millman.
学生们似乎无法抗拒大型语言模型为他们带来的便利，这些模型帮助他们整理和塑造想法，而当语言可以被机器生成时，他们便无法忍受寻找语言的痛苦。
Students seem unable to resist the ease that large language models provide them to wrestle with and craft their ideas and are incapable of tolerating the discomfort of searching for language when language can be produced for them.
他们似乎也不愿意在困难中多待一会儿，让困难教会他们一些东西。
They also seem unwilling to stay inside difficulty long enough for the difficulty to teach them something.
使用人工智能有点像吸毒，不是因为它带来快乐，而是因为它消除了某种痛苦。
There's something drug like about using AI, not because it produces pleasure, but because it removes a particular kind of pain.
然后她有一句很棒的话。
And then she has this great quote.
在一个人让机器替自己写作之前，他们应该知道从自己的不确定性和身份深处写作是什么感觉。
Before a person asks a machine to write for them, they should know what it feels like to write from the depths of their own uncertainty and identity.
在一个人让机器模仿自己的声音之前，他们应该经历过获得自己声音的那个漫长、曲折、常常令人难堪的过程。
Before a person lets a machine imitate their voice, they should have endured the long, uneven, often humiliating process of acquiring one.
我很喜欢这句话。
I love that.
是的。
Yeah.
展开剩余字幕（还有 480 条）
我太喜欢这句话了。
I absolutely love that.
我觉得这就是为什么对我来说，汉克没问题。
And I think that's why Hank to me is fine.
但我理解，年轻人可能没有机会在思想被轻易代劳之前，充分发展自己的思考和世界观。
But I understand where younger people have not had the opportunity potentially to fully develop their own thoughts and world without the ease of a thought being developed for them.
我很快要联系一个你可能觉得疯狂的事情。
I'm gonna make a relationship to something that you may think is crazy really quickly.
肝脏之王。
Liver King.
还记得那个人吗？
Remember that guy?
你说得对。
You're right.
我确实觉得那很疯狂。
I do think that's crazy.
你还记得吗，大家都看到他的腹肌了？
You remember, like, everyone saw his abs?
对。
Yes.
他的腹肌简直了，哇。
And his abs were like, woah.
他用了这个方法，才练出那样的腹肌。
He does this method, and that's how he gets them.
然后真相曝光了。
And then it came out.
他一直在吃类固醇。
He was taking steroids.
是的。
Yeah.
对我来说，这感觉就像是一个很类似的对话，你会说，哦，对，那不是你靠自己得来的。
I to me, that feels like a very equivalent conversation that, like, you're like, oh, yeah, you didn't earn that.
在这里，产生一个原创的想法本身就是件很了不起的事。
Earning an original thought is a fascinating thing here.
这其实就是我们讨论的核心。
That is actually what this is about.
对吧？
Right?
任何原创的想法都必须经过努力才能得来。
Any original thought needs to have been earned.
我同意这一点，而且我百分之百相信，我们之所以有这种关系，是因为在没有这项技术的情况下，我们已经做了三十七年的创作者。
I agree with that, and I a 100% am in on maybe we have this relationship because we've been creators for thirty seven years without this technology.
我用VHS摄像机拍过东西。
I shot with VHS cameras.
我写日记。
I write in a journal.
我知道怎么去做那些传统的事情。
I know how to do the other thing.
对。
Yeah.
而且我经历过那种痛苦。
And I've gone through the pain that is that other thing.
所以现在我看到什么东西，能立刻判断好坏，因为我知道它是否真的包含了艺术创作的过程。
So I now when I see something, I can immediately assign good, bad to it because I know if it actually includes the artistic process or not.
让我对汉克那件事的反弹感到失望的是，人们想要他诚实。
The thing that gives me like, that made me disappointed about the backlash to Hank's situation was that people wanted honesty from him.
然后他给了他们。
And then he gave it to them.
他非常坦诚地说，
He said very candidly
对。
Yeah.
这就是我在做的事。
This is what I'm doing.
如果这不是你想要的，我很抱歉。
I'm so sorry if this is not what you wanted from me.
如果让你失望了，我很抱歉。
I'm sorry if I let you down.
这就是我在创作过程中使用它的方式。
This is how I use it in my process.
他详细说明了Complexly明确的AI政策，这个政策已经实施三年多了。
He outlined Complexly's very clear AI policy that has been in place for over three years.
这不是昨天才制定的。
This is not something that was made yesterday.
我仔细读过了。
I read through it.
他们从来不用它来生成内容。
They never use it to generate Yeah.
比如他们的脚本，他们从来不用它来生成视频或图片。
Anything, Like their scripts, they never use it to generate video or images from Yep.
未经同意的LLM，也就是那些公开的聊天机器人，比如ChatGPT、Claude、Gemini。
Non consensual LLMs, meaning any like public one chat, GBT, Claude, Gemini.
对。
Yep.
他们是在一个封闭的循环里使用它，和Yep一起。
They use it on like a closed loop with Yep.
经过同意的图像。
Consensual images.
但反对的声音依然存在。
And the backlash still persisted.
这让我真的很失望，因为你希望你的创作者是诚实的。
And that makes me really disappointed because you want honesty from your creator.
你说得对。
You got it.
我觉得那时候你可以选择是否继续关注他们的内容，如果那个政策符合你的标准的话。
And I think you could choose at that point to engage or not engage with their content if that policy is up to your standards.
但我认为，在Hank出面解释了他的立场之后，大家还继续对他进行抨击，这完全没有必要。
But I think the continued backlash against Hank after he's come through and explained his side of things is completely uncalled for.
大家就是完全不愿意去理解其中的细微差别。
There's there's been like a complete unwillingness to engage in the nuance of
对。
this.
没错。
Exactly.
我觉得这反而给我们带来了一个机会。
I think that it it opens us up to an opportunity.
比如，读一下Complexly的AI政策，我觉得真的很有意思。
Like, reading Complexly's AI policy, actually thought was really interesting.
它非常简单。
It's very simple.
就一页纸。
It's one page.
我们会把链接放出来，你们可以自己看，或者直接搜一下，Complexly 的 AI 政策。
We'll link it so you guys can read or just look it up, Complexly AI policy.
基本上就是一个谷歌文档，里面列了大概四条规则。
It's basically a Google Doc that maps out like four rules.
我觉得如果能展开一个健康的讨论会很好，我们在这里先起个头，但我觉得应该放到评论区去。
And I think what would be a really healthy discussion, and we'll start it here, but I think it should take place in the comments.
这应该是创作者之间的一场讨论，就是：我们的 YouTube 创作者 AI 政策应该是什么？我们应该共同制定它吗？
It should take place as a discourse amongst creators, is what is our YouTube creator or creator AI policy, and we should collectively create it?
我们未来使用或不使用 AI 的原则是什么？
What are the principles in which we will be using or not using AI moving forward?
我抛个砖吧。
I'll throw out a line.
好。
Okay.
这条线就是：观众能察觉到的地方。
The line is wherever the audience notices.
那是你写的吗？
Did you write that?
对。
Yeah.
你觉得我喜欢用L来写那个吗？
You think I like using L underwrite that?
我不能
I can't
打印任何东西你
print How anything you
我们在给它评级吗？
are we ranking it?
我们要给绿色吗？不，不。
Are we going to go green, No, no.
黄色，
Yellow,
所以我觉得，这个挺有意思的。
So I mean, that's an interesting one right there.
再说一遍。
Line say it again.
只要观众看不出来，那就没问题。
Line is wherever the audience notices.
不。
No.
我觉得这个功能上的理解是对的，因为我们可以讨论理论上是否合适。
I think that's the correct functional take because we can talk about, like, theoretical whether it's fine or not.
但如果人们没注意到，那就没问题。
But if people don't notice, then it's fine.
实际上就是没问题，因为我们对所有YouTuber的了解就是这样。
Like, it's effectively fine because that's, like, what we know of all of the YouTubers.
那是不是意味着，如果我用它来写我的脚本，没人知道，那就在你刚才设定的范围内？
Does that mean if I use it to write my script and no one knows, then that's within your parameter that you just set?
他就是这个意思。
That's what he's saying.
对。
Yeah.
黄色。
Yellow.
偏红的黄色。
Yellow leaning red on that.
对我来说，关键是你用AI做的事情，是不是你团队里本来有人能做的？
For me, it's about are you using AI for something that someone on your team could be doing?
所以道德问题在于人力资源。
So the morality is around, human resource.
对。
Yeah.
比如说，我有一段音频效果很差，我想用Adobe Podcast来修复它。
Let's say I I have, like, audio that's messed up and I want to use Adobe Podcast to fix it.
我不知道怎么修那段音频。
I don't know how to fix that audio.
这个AI能修好音频。
This AI can fix the audio.
我现在把一项技能外包给了机器。
I've now outsourced a skill to a machine.
我猜我指的是生成式AI。
I guess I meant in like the generative AI.
那是生成式AI吗？
Is that generative AI?
那是生成式的。
That's generative.
对。
Yeah.
我不知道。
I don't know.
对。
Yeah.
我觉得那算是屏幕吧，对吧？
I think So that's screen, right?
因为比如说你是个小创作者。
Because let's say you're a smaller creator.
你不会剪辑。
You don't know how to edit.
你的麦克风很烂。
You have a crappy mic.
你买不起400美元的麦克风。
You can't afford a $400 mic.
它能让声音更好听，相比之下，缩略图呢。
It makes it sound better, Versus a thumbnail.
比如说你有个缩略图，里面有个很精致的细节，很有艺术感的细节。
Let's say you have a thumbnail and there's, like, a really neat detail in it, like, really artistic detail.
如果你是随便刷视频的人，你会觉得，哇。
And if you're a scroller, you'd be like, wow.
就像，有人特意把它放在那里。
Like, somebody put that there.
就像，我在想这幅画里它代表什么。
Like, I wonder what that represents in this art.
我有一个
I have a
这挺有意思的。
That's interesting.
我有个后续问题，因为你一开始提到小创作者，那个成本问题确实很有意思。
I have a follow-up there because you started that with saying you're a small creator, and that affordability aspect is interesting.
那如果是大创作者，或者比如说，一个YouTuber给视频配音，跟大型Netflix制作用这个来配外语，情况会不一样吗？
Why does that does it change if you're a bigger creator or, like, let's say, a YouTuber dubbing videos versus a large Netflix production using it to to dub in a foreign language?
这个界限在哪里？
Where's the line there?
我觉得首先是个努力程度的问题，预算也是一部分，但就像如果你是MrBeast，你需要做几十个缩略图，对吧？
I think there's a question of effort for one, and there's also like, budget is part of it, but it's like, if you're mister beast and you have, like, dozens of thumbnails that you need to be you need made, right?
他每个视频大概要做100个。
He makes like 100 per video.
一般来说，AI做的缩略图不如人工做的好。
Generally, an AI thumbnail is not as good as a man made thumbnail.
大概只有85%的水平。
It's generally 85%.
你可以花点功夫把它提升到更高的水平。
You can put that effort to get to that next level.
而且，当有人类参与制作时，我觉得更有可能创造出新的东西，而且有人会……
And also, when there's a human making it, I think there's more odds of you making something new and someone's
我们又回到了你最初提出的那个框架，就是道德问题。
We're back to the original frame that you put forward is morality.
所以我要再描述一个现实生活中的场景。
So I'm going to paint one more picture here of a real life scenario.
很长一段时间里，我们和Eleven Labs合作，给我们的视频配音，那是一家AI语音公司。
For a long period of time, we dubbed our videos in partnership with Eleven Labs, which is an AI voice company.
我们把视频配音成了多种语言。
We dubbed them into multiple languages.
我们本来想像Beast那样，用真人演员来配音。
We wanted to do that with, human actors the way that Beast did it.
我们拿到了一份报价，因为我们有个视频差不多两小时长，配成10种语言要23万美元，这可不是一笔小数目。
We got a quote for one of our videos because it was almost two hours long for $230,000 to dub it into, 10 languages, which is a lot of money.
那些可都是真人配音演员。
Those are real human voice actors.
我们最后的选择是用Eleven Labs，用AI语音把视频配成所有那些语言。
The choice we made was to use 11 labs to dub it into all of those languages through AI voice.
对。
Yeah.
那Eleven Labs花了多少钱？
I mean, how much was 11 labs?
那简直是小巫见大巫。
I mean, nominal compared.
具体价格记不清了，但也就几百美元。
Don't remember the price, but in the hundreds of dollars.
是啊，我觉得AI配音那点钱，根本不算什么，对吧？
I mean, yeah, I think the AI dubbing, like that's bread, right?
那不就是小钱吗？
That's bread?
那确实是小钱，对吧？
That is bread, right?
但问题是——等等，等等，那是大钱？
But the thing is, it's like- Wait, wait, that's red?
大钱？
Red?
小钱，小钱，小钱。
Bread, bread, bread.
就像面包一样。
Like it's bread.
很贵。
It's expensive.
就是很多钱。哦，对，就是。
Like it's a lot Oh, of it's
面包。
bread.
哇。
Wow.
好。
Okay.
明白了。
Got it.
对，对，我懂了。
Yeah, yeah, I see.
对。
Yeah.
对。
Yeah.
所以我觉得，首先，如果你沿着这个谱系往下看，到某个点听起来就像我们在拼命硬撑。
So I think for one, right, if you go down the spectrum, it's going to sound like we're coping really hard at a certain point.
就是这样，对吧？
Like that's just how it Right?
我也用过AI。
I've used AI too.
对吧？
Right?
我们都是这么运作的。
It's how we all function.
我觉得，当它被用来做那些你知道那个人本来可以自己做的东西时，就像你是个小孩一样。
I think it's just like when it's used for something that you know the person could have made themselves, it's like you are a child.
你就是懒。
Just you are lazy.
你就是个失败者。
You are a loser.
如果你自己能做，那就自己做。
Just make it yourself if you can make it yourself.
之所以没有细微差别，就是因为没人去想这些。
It's like there isn't the reason there's no nuance take is because nobody is thinking about That's
是尊重，不是道德。
respect, not morality.
对吧？
Right?
就像，那是艺术上的尊重，我同意这一点。
Like, that's artistic respect, which I agree with.
但道德是另一回事。
But morality is a different thing.
对。
Yeah.
而且我，我再说一遍，我是在推动对话向前走，不是要……
And and I'm I'm again, I'm I'm pushing to push the conversation forward, not to.
我很感激你的推动，因为我没法把每个音频文件都转录出来。
And I I appreciate the push because I I can't transcribe every audio file that I do
以后不能再帮你转录了。
for you anymore.
你为什么不呢？
You don't why?
你刚才说你很感激。
You just said I appreciate it.
是的。
Yeah.
你从哪儿听来那句话的？
Where'd you get that line?
我说了吗？
Did I?
说了。
Did.
我在开玩笑吗？
Am I kidding?
完全信任你。
Trust you at all.
对。
Yeah.
对。
Yeah.
那太疯狂了。
That's crazy.
你说我在反驳，然后
You said I'm pushing back, and
我很欣赏你的剧本，Sid。
I appreciate script, Sid.
对。
Yeah.
对。
Yeah.
这简直就像个小Sid GPT。
There's It's literally little Sid GPT.
我的天。
Holy shit.
好。
Okay.
所以我知道，Connor，你有一些情景。
So I know, Connor, you have some scenarios.
是的。
Yeah.
我确实有一些场景。
So I have some scenarios.
我觉得最有帮助的做法是，我们逐个过一遍，讨论一下某个场景是否合适。
I think what would be super helpful is we just go through, talk about whether or not we think a specific scenario is okay.
希望到最后，我们能总结出三个原则。
And hopefully by the end, we can come up with three principles
好。
k.
这些原则是我们都认同的，关于创作者如何以自己觉得舒服的方式使用AI，虽然我不确定用“道德”这个词是否恰当。
That we agree on for I don't know if ethical is the right word, but how creators can go about using AI in a way that they feel good about it.
好的。
Okay.
创作者用AI生成缩略图的背景和人物，但用的是自己的脸，然后手动用Photoshop把脸P回去。
Creator makes a generated thumbnail background and person, but they use their own face and manually Photoshop it back on.
绿灯、红灯还是黄灯？
Green, red, or yellow?
红色。
Red.
黄色。
Yellow.
我这边完全是绿色。
I'm full green here.
对。
Yeah.
对我来说，我知道很多我尊重的人和创作者都在大规模这么做。
I think for me, I mean, I know this is happening at scale with a lot of people and creators I respect.
而且我觉得，他们这么做还是有一定水平的，我也说不好。
And I think, there's still a level of, I don't know, they're doing it.
他们在构思这个。
They're concepting it.
我看着他们做。
I'm watching them do it.
我觉得只要图片，这更多是个偏好问题，只要缩略图和图片能准确反映视频内容，我就没问题。
I think as long as the image, and this is just more of a preference thing, as long as the thumbnail and the image accurately represents what's inside the video, I have no problem.
我觉得我们总体上在往一个方向发展，即使作为观众，我也更喜欢真实照片。
I think generally we're moving in a direction where even me as an audience member, I prefer like a photo.
我们最近聊得比较多的是，我想要更朴素的缩略图，因为谁都能做那种花哨的。
We've been talking about that a lot more is like, I want a lower FI thumbnail because everyone can do that.
现在看起来都千篇一律，就像你和我。
It now looks homogenous like you I
我不想看到缩略图时还要问，这环境是真的吗？
don't wanna have to ask the question when I see a thumbnail, is that the real environment?
对。
Yeah.
所以对我来说，界限就是观众能不能看出来。
So that to me again is like, the line is wherever the audience notices.
没错。
Right.
我觉得如果
I think if
我没注意到，点了进去，视频内容也跟得上，那我就没问题。
I don't notice and I click and the video backs it up, I have no problem.
不过说到缩略图，你想想看，它们其实已经假了很久了。
But with thumbnails, if you think about it though, they have been looking fake for a while.
就像人们一直在修图让它看起来假，这挺有意思的。
Like people have been editing to look fake, which is just, it's interesting.
有意思的是大家对此意见不一。
It's interesting how people are torn on it.
而且，我们俩都非常尊重缩略图设计师。
Also, like both of us have the utmost respect for thumbnail designers.
我是说，我们雇了
I mean, we employ
我们花了不少钱
We some a lot of money
用于缩略图设计。
for thumbnail designs.
而且说清楚，我们
And to be clear, we
很大程度上是付钱请他们做策略的。
pay them partly in large part for their strategy.
对。
Yep.
我们也知道，他们用AI工具做缩略图，比我们自己去用这些工具做得更好。
Knowing also that they are using AI tools to make thumbnails for us better than we could use those tools.
如果有人用AI来生成视频大纲，然后自己再去填充内容或者写脚本，那会怎么样？
What if someone uses AI to create an outline for a video that they go and fill out or script themselves?
红色。
Red.
黄色。
Yellow.
我这边完全是绿灯。
I'm fully green here.
我觉得拿个大纲，这就是我常做的事，我会问，你知道的，你会怎么处理这个？
I think getting an outline, this is something I do, I ask, you know, how would you approach this?
我的故事里有哪些漏洞？
What are the holes in my story?
对，这就是我常做的。
Like, that is something I do.
我来这儿之前的工作就是给评论类YouTube频道做研究、列大纲、写脚本。
That was my job before I got here is I was a researcher, outliner, scripter for commentary YouTube channels.
而且那是个挺冷门的工作。
And it was a very like sparse job.
就是没多少人干这个。
Like not a lot of people had that job.
嗯。
Mhmm.
那是在ChatGPT出现之前，我当时真的很感激，因为正是那份工作让我今天能坐在这里。
I was this was like pre chat GBT and I was like really grateful because that's that's how I ended up here in this room.
对。
Right.
而且我觉得这件事其实有人能做，虽然我当时工资不高。
And I think that is something that someone could do for I was not being paid a lot for that.
非常便宜。
Very cheap.
而且花的时间也不长。
And didn't take me very long.
是的。
Yeah.
但如果做得不好，那结果就是不好。
But isn't isn't it just like the but if that's bad, it's bad.
比如你现在上领英，大概75%的内容都是大语言模型写的，而且质量很差。
Like, if you go on LinkedIn right now, like, 75% of the the content is written by an LLM and it's bad.
而且其中70%都是你发的。
And 70% are yours.
这其实也是我
It's actually why I
发帖这么少的原因，因为我看到那个环境，我就觉得，我得想出一个原创的观点，才能给这个生态添点东西。
post so little because I'm so, I look at that environment and I'm like, I have to work on an original thought and take to add to this ecosystem.
而很多人只是在那儿拼命发帖，但内容都很差。
And so many people are just cranking posts out there, but they're bad.
读起来一点意思都没有。
They're not fun to read.
那些人没有自己的品味，不管有没有用大语言模型。
Those people don't have their own personal taste, LLM aside or not.
对。
Yeah.
或者说他们的品味跟你不合。
Or they don't have a taste that aligns with yours.
没有品味，你的内容读起来就没意思。
Don't have taste makes your it
读起来没意思。
uninteresting to read.
那我们试着把这些都整合起来，一起头脑风暴一下。
So let's try and bring this all together and brainstorm.
在我们思考如何使用AI，或者其他创作者如何在创作过程中使用AI时，有没有三条清晰的原则或规则可以遵循？
Are there three clear principles or rules that we can follow as we're thinking about how we use AI or how other creators can use AI in their process?
我们先从构思开始吧。
Let's start with, like, ideation.
对。
Yeah.
头脑风暴。
Brainstorming.
比如说，我想做一个关于Hank Green的视频。
Let's say, you know, I'm trying to make a video about Hank Green.
这个视频，我需要为这期节目准备100个标题选项。
This video, I need a 100 title options for for this this episode.
对。
Yeah.
所以在我处理大多数悲剧性内容时，它的用法就是帮你开个头。
So the way I've used it in my Like for most tragedy things, it can help you get the ball rolling.
如果你让它生成500个标题，你可能会得到一个勉强能用的。
If you ask it to generate 500 titles, you will probably get one that is half usable.
但这是个给你灵感的好方法。
But it's a good way to give you ideas.
所以就像，我在这里需要一个好的形容词。
So it's like, I need a good adjective here.
那你有哪些形容词？
What that's are your adjectives?
构思。
Ideation.
我们能把它简化成什么呢？
What can we boil that down into like
那是什么？
What what is that?
对。
Yeah.
那是什么？
What is What is that?
核心原则是什么？
What's the principle?
是不是说，AI作为辅助角色是可以的？
Is it like AI as supporting character is okay?
我觉得这里的一个问题是，对我们很多在YouTube上长大的人来说，它的美妙之处在于，一个人在自己的卧室里，对某件事充满热情，他们唯一的资源就是自己、一台摄像机和自己的想法，他们会想尽一切办法把内容发布到世界上。
I I think one of the issues here is that like, for a lot of us that grew up on YouTube, like the beauty of it is an individual out of their bedroom who cares so deeply about something, and the only resource they have is themselves and a camera and their own thoughts, and they will do whatever it takes to get it out into the world.
而现在，这种情况正在发生巨大的变化。
And now that is just changing drastically.
现在有些东西已经不一样了——
Like there is something now-
这种变化是通过YouTube主流化、商品化的视角发生的。
That has been changing through the lens of the mainstreamification of YouTube, the commodification
完全同意，不管有没有AI，这都在发生。
Totally, that's happening with AI or not.
Mr. Beast的存在本身就说明了一切。
The pure presence of a Mr.
他的出现就是时代变化的标志。
Beast is a sign of how things have changed
看看我们，我们当初就是两个在你车里拍视频的人。
Look since at us, we were two guys who made videos in your car.
现在我们有了团队。
We now have a team.
我们雇了一群人。
We employ a bunch of people.
我们还在不断招人。
We are continuously hiring more people.
我们现在是在经营一家公司了。
We run a business now.
我觉得大家对YouTube的那种情绪，可能也有一部分是因为我们正在见证从个人创作者变成媒体公司的转变，因为这种转变会推动效率的提升，而效率的提升又会促使内容被产品化。
And I think that is also probably some of the emotional feeling towards YouTube is the change from individual to media company that we're watching at the same time, because that encourages a higher level of efficiency, which then in turn encourages a productization of something.
说到效率，回到Hank Green那个情况，其实很多问题就出在效率上。
And efficiency, getting back to the Hank Green situation, is a lot of the problem, I
对，没错。
think Yes.
对他的很多观众来说。
For a lot of his audience.
突然之间，他做视频的方式让人觉得，视频就像是在清单上打了个勾。
That all of the sudden, the methods that he used to create the video make the video feel like a checked box on a checklist.
对。
Right.
他是在打勾完成任务，而不是出于内心深处的分享欲望。
That he's checking off instead of a deep rooted personal desire to share.
所以对我来说，这是第一原则。
So that to me is number one principle.
如果我们要总结一下，对我来说，真诚才是YouTube上、文化中、以及各个领域里能够脱颖而出、引起关注的东西。
If we were to boil it down, is honesty for me is what is breaking through and what's interesting on YouTube, in culture, across the board.
要真诚。
Be honest.
真诚可以是关于你的使用情况，但也可以是关于做你想做的东西。
And being honest can be like being honest about your use, but being honest also about making something you want to make.
我觉得这是因为，如果我真的想做某件事，并且得到了在座各位的帮助，再加上通过大语言模型整理思路，最后用我真诚的视角呈现出来。
And I think that's like because if it's something I really want to make and I get help from people in this room plus organize those thoughts through an LLM and then deliver it through my honest lens.
对我来说，如果我是接收方，我是可以接受的。
To me, if I'm on the receiving end of that, I'm okay with it.
如果你给我的是你真心想做的内容，我完全没问题。
I'm okay if you're delivering me something that you honestly want to make.
我完全同意。
I completely agree.
我觉得尽可能多地披露，只要不泄露你的商业机密之类的。
I think disclosure as much as you can, like, without giving away your business secrets or whatever.
但我认为，尽可能向观众披露你使用了多少AI以及何时使用，这真的很重要，因为这样观众就能根据他们听到的内容来决定是否想继续关注你。
But I think disclosing to your audience how much AI you're using and when whenever possible is really important because then your audience can decide, do I want to engage with this person based on what I'm hearing?
我不认为这会改变你的观众数量。
I don't think that would change your viewership at all.
而且你对自己如何使用AI非常坦诚。
And you are very honest about how you use AI.
不过，如果我正在看的创作者说，顺便提一下，我用了带AI的调色软件来指导我调色，我会觉得有点奇怪。
I would almost feel bizarre, though, if a creator I'm watching is like, by the way, I used a color grading software that used AI to tell me how to color grade this.
我会觉得，
Like, I would be like
我觉得这其实就是个习惯问题。
I think it's just norms.
对吧？
Like, you know?
咱们看视频吧。
Let watch the video.
你知道Renegade在TikTok上火了，但大家都没给原创者署名，对吧？
You know when Renegade went on TikTok and everyone, like, did not credit the original Yes.
我觉得TikTok上后来形成的舞蹈署名规范，类似的事情也可能发生。
I think the DC, like the dance creds thing that emerged as a norm on TikTok, I could see something similar happening like
好。
Okay.
比如LLM署名，或者像视频里用了什么工具这样的标注。
Like LLM creds or like where like what tools did I use in this video?
我觉得类似的东西可能会出现。
I could see something similar to that emerging.
不过我们可能不会列出来。
Now, we're not gonna probably list off that.
我们大概不会在页脚写上“由Riverside录制”之类的。
Like, we're probably not gonna put in the footer like Riverside used for Sure.
每期新闻通讯的转录也不会写。
Transcription of every newsletter.
但如果缩略图是用Nano Banana做的，我觉得还是应该标注一下。
But I think when it's like thumbnail made in part by Nano Banana.
那你觉得我用Photoshop的时候也要标注吗？
But do you want me to put that for Photoshop too?
还是说用Photoshop的话，你就不指望我标注了？
Or your expectation wouldn't be there if I was using Photoshop?
所以我才说，能标注就尽量标注。
This is why I'm like disclosure when you can.
或者说，在你觉得有必要的时候标注。
I or like disclosure when you feel like it's necessary.
所以，说实话，这整期节目最后可能就归结为“看情况”这三个字。
And that's why honestly, unfortunately, the whole the whole episode is about to boil down to it depends.
对。
Right.
看情况。
It depends.
所以这就是答案。
So that that's the answer.
这个答案，
Which the answer.
对我来说，
Which to me,
我觉得是个好结果。
I think is a good outcome.
是的。
Yeah.
是的。
Yeah.
如果我们能达成共识，这里面有细微差别，而且这是新事物，我们都在摸索中。
If we can agree that there is a level of nuance and this is new and we're all figuring this out together.
作为创作者，如果，再说回Hank，我对这个主题的顾虑在于，当道德被强加于其上时。
And as creators, if we, again, like with Hank, what my issue with the subject matter is, is when morality is placed on it.
道德和那种感觉，对我来说，和‘是的，我在做某事’是两回事。
Morality and like that to me feels like a different thing than like, Yeah, I'm doing something.
我做得太多了。
I'm doing too much.
对吧？
Right?
因为我觉得，是的，突然这些工具让我们能做更多，但如果我们做得更多，如果我们不是通过诚实来驱动效率，我们可能都在做出更差的东西。
Because I think it's like, Yeah, all of a sudden these tools are allowing us to do more, but we're all probably making worse stuff if we're doing more, if we're driving efficiency not through our honesty.
那只会是更差的东西。
It's just gonna be worse stuff.
所以让我们回到我们能做的、好的、诚实的事情上，因为那总是能突破，但我们正处于一个全新的时代。
So let's pull back to like what we can do that's good and honest because that has always broken through, but we're in a totally new era.
所以我们正在一起摸索。
And so we're figuring it out together.
我觉得观众和正在看、正在听的人，也在和我们一起参与这个过程。
And I think the audience and people who are watching, listening right now play a role in figuring this out with us.
但说到恶意意图，或者假设别人有恶意，我就觉得，我们认识很多YouTube上的顶级创作者。
And the malicious intent or the assumption of malicious intent is where I'm like, that to me, we know many of the top creators on YouTube.
我很难站在这里说，他们在使用技术时带有恶意或不道德。
It is very hard for me to stand here and go, They have malicious intent or some immorality in their use of technology.
不过我觉得，对于整个AI的问题，我的反驳是，你能看出来吗？
I think though, my counter to this whole like AI, like, can you tell?
你看不出来吗？
Can you not tell?
我想问一下Samir和Do Collin，你们不觉得，当你看到一个视频，屏幕上首先出现的是“这个视频里的所有素材都不是用AI制作的”，这很酷吗？
For I I want to ask both Samir and Do Collin you guys not think that it is aura when you watch a video and the first thing that comes up on the screen is none of the assets in this video were made using AI.
这难道不会让你觉得，哇，这就像你在看……
Does that not make you feel like, yo, this this is like you're watching
这是Z世代的事。
Gen Z thing.
那不可能。
So no way.
这是代际差异，因为我有这种感觉。
That's a generational thing because I feel that.
所以我要说，Kish，对我来说，那是一种讲故事的手法。
So I will say, Kish, like, for me, personally, that is a storytelling device.
对。
Yes.
就是，
Like,
你知道，Andreas Hem，一位了不起的电影制作人，他拍了一个很棒的视频，里面就用了这个手法。
you know, Andreas Hem, incredible filmmaker, made a great video where he did that.
他说这里面没有任何东西是用AI做的。
He said none of this is made with AI.
我现在就用这个视角去欣赏那个视频，看他拍的每一个镜头，感觉完全不一样了。
I am now using that as a lens to enjoy the video in a different way as I see the shots that he made.
那如果娜塔莉·林恩的视频是用AI做的，你还会有一模一样的感受吗？
So would you still feel the same things you feel when you watch a Natalie Lynn video if it was made with AI?
因为那些视频大多都非常个人化。
Because most of those videos are so personal.
或者凯西，不，但我个人觉得
Or Casey, no, but I don't
那是不可能的。
think that's possible personally.
如果有一天它变成可能了，而我还没察觉，那我不知道，但就目前而言，我不相信能做到。
If it's becoming possible and I don't notice, I don't know, but that presently I don't believe is possible.
娜塔莉身上有的是真诚。
That what Natalie has is honesty.
她在讲一个真实的故事。
She's telling an honest story.
那么，五年或三年后，我们会达到那个程度吗？
Now, in five years or three years, do we get to a point?
也许吧。
Maybe.
但我觉得作为人类，我们现在对一个人讲的故事是否真诚有很好的辨别力。
But I think as humans, we have a really good detection right now of when someone's telling an honest story or not.
我就是不相信这能做到。
I just don't believe it's possible.
给我看一个能带来那种感觉的视频。
Show me a video that feels like that.
嗯，问题就在这里。
Well, that's the thing.
你不会知道的。
You wouldn't know.
你看过一个视频，你也不会知道。
You wouldn't know if you have watched a video.
对。
Yeah.
我的观点是
My argument
然后我觉得在最后，我依然感受到了同样的美和惊奇。
And is then I think I still felt the same sense of beauty and wonder at the end of the If
你，对。
you Yeah.
不过，那不一样。
It's not the same, though.
就像，只是
Like, just
太棒了。
amazing.
我想把话题稍微拉回到原则上来。
I'm I'm gonna bring us back into the principles a little bit.
是的。
Yes.
所以第一点，我们要诚实，既要说明你如何使用AI，也要做你真正想做的事情。
So number one, we have be honest in both how you're using AI and making something you wanna make.
我觉得第二条原则，就像你说的，应该让它感觉像人，甚至难以分辨。
I think the second principle is along what you said that, like, it should feel human to the point where it's indistinguishable.
对我来说，这一点很重要。如果我知道或者不知道它是用AI做的，体验都一样的话，我一般都能接受。
Like, I think that is a big thing for me, whereas, like, if I can have the same experience with or without knowing that it was made with AI, I think, generally, I'm okay with it.
我这么说是因为，根据我的经验，和Samir类似，我也尝试过做早期采用者，用AI做各种事情。
And I only say that because like, from my experience, I have similar to Samir, I've tried to be an early adopter, and I've used AI to do different things.
我用AI为今年我们和Speed的James一起做的第一个视频制作了图形。
I used it to make graphics for the first video we made this year with James from Speed.
你知道，我是个做图形的人。
And, you know, I'm someone who makes graphics.
我会在After Effects里做。
I'll make them in After Effects.
所以我的标准是，我能不能把它做得像我想要的那样好，不管用什么工具？
So my line was, Can I make this good as Can I make this the way I wanna make it regardless of the tool?
然后那个视频里有一个图形在评论区被点名了好几次。
And there was one graphic that got called out in that video in the comments a couple times.
我就说，哦，好吧，这里有个界限。
And I said, Oh, okay, there's a line.
我还没能熟练到这个工具，做到让我的观众看不出来。
Am not yet good enough at using this tool to make it indistinguishable for my audience.
他们注意到有点不对劲。
They are noticing that something is up.
对我来说，那就是界限。
And that to me was the line.
不是说等我更熟练了就不会再用它了。
Not to say that I won't use it again when I get better.
但我觉得，目前这就是我的界限，因为我还没达到应有的水平，确保它不影响故事。
But I was like, Currently, that's my line because I'm not as good as I should have been to make sure it didn't impact the story.
对。
Yeah.
我觉得还有一种真诚的兴奋感。
I think there's a level of honest excitement too.
而且我们得展示出来，比如我们最近建了好几个网站和定制应用，我觉得都特别酷。
And we have to put forward, like, we've built multiple sites recently and custom apps that I think are really cool.
我们为雷克萨斯活动做的那个网站，鼠标指针就是一辆雷克萨斯车在屏幕上跑来跑去。
Our site for a Lexus campaign has the Lexus car driving around as the cursor.
那是用AI做的。
That's built with AI.
我们刚为频道做了一个测验，全部基于Gemini构建。
We just built a quiz for our channel, and that's all built on Gemini.
而且特别酷。
And it's super cool.
我真的很兴奋能把这些东西展示给大家。
And I have honest excitement to show this stuff to people.
而且我觉得这也是一种层面，就是关于Hank发生的事情，让人觉得他在背后偷偷做些什么，自己都感到羞愧。
And I think that's a level too where it's like there's something about what happened with Hank that feels like he's doing something behind closed doors that he's ashamed of.
对吧？
Right?
而且这对他来说也是真实的。
And that's honest to him.
对，就是这样。
And that's yeah.
而且这对他来说也是真实的。
And that's honest to him.
对。
Right.
可能
There could
会有一种情况，Hank说，我在用这些东西，而且我会继续用下去。
have been a scenario where Hank said, I am using these and I'm gonna continue to use them.
完全同意。
Totally.
抱歉，你注意到了这一点，而且感觉不太对劲。
Sorry that, like, you noticed this one and it didn't feel
对。
Yeah.
我能有这种感觉吗？
Can I feel right?
从品牌的角度问一个简单的问题。
A quick question from a brand perspective.
如果他说的是Claude或Gemini，世界的反应会不一样吗？
If if he said Claude or Gemini, would the world have a different reaction?
观众对Chatuchu BTS的反应会不同吗？
Does the audience have a different reaction than Chatuchu BTS?
是的。
Yes.
是的。
Yes.
我觉得如果他说的是Gemini，可能不会。
I think if he said Gemini, maybe not.
如果他说的是Claude，那绝对不一样。
If he said Claude, absolutely different.
我觉得这对我来说特别有意思，抱歉。
I think it is so interesting to me to see sorry.
我刚才说什么来着？没有。
What did I No.
没有。
No.
没有。
No.
继续说。
Keep going.
对。
Yeah.
确实很有意思。
It's fascinating.
我觉得特别有意思的是，这些公司做的事情几乎一模一样，对环境的影响也差不多。
It is so interesting to me to see these companies that are doing pretty much identical things having identical environmental impacts.
嗯。
Mhmm.
其中一家在免费版里不投放广告，比如我知道Claude就没有广告。
One of them doesn't place ads in its like free like, I know Claude doesn't have ads.
这是他们的特色，在回复里也是这样。
That's their whole thing like in their prompted responses.
据我所知，免费的ChatGPT是有广告的，如果我没记错的话。
I know that the free chat GBT does, if I'm not mistaken.
除此之外，它们几乎没什么区别。
Besides that, they are pretty indistinguishable.
但ChatGPT却承受了更多的批评，我一直不明白为什么我们没有用同样的态度去对待所有这些大语言模型。
And yet chat GBT takes so much more flack, and I have never understood why we are not treating all of these LLMs with the same amount of disdain.
我觉得部分原因是ChatGPT是一个更明显的代表人物，就像Sam，Sam。
I think I I think part of it is Chad GPT is a much clearer figurehead in Sam Sam.
对。
Yeah.
攻击一个代表人物要容易得多，而像Anthropic的负责人是谁我都不知道。
Which is way easier to attack a figurehead, whereas, like, I don't even know who the head of Anthropic is.
真的吗？
Really?
但是
But
对。
Yeah.
昨天刚听到他的名字，但我记不住。
Just heard his name yesterday, and I can't remember it.
达里奥，我觉得提这个挺有意思的。
Dario, I do think it's interesting to bring up.
我是说，我永远忘不了Anthropic退出国防部合同那件事。
I mean, I'll never forget when Anthropic pulled out of the contract with the defense department.
对。
Yeah.
然后OpenAI接手了那个合同，大家都说，我们要抛弃OpenAI。
And OpenAI stepped in and got the contract, and everyone was like, we are dropping OpenAI.
我们要抛弃ChatGPT。
We're dropping ChatGPT.
大家都去下载Claude吧。
Everyone download Claude.
我就想，如果条件合适，Claude难道不会接那个单吗？
I'm like, you think Claude wouldn't have taken that deal if the deal was right?
就是，对。
Like Yeah.
他们当时肯定也会跟着他一起的。
They would have been right there with him.
咱们把话题拉回来吧。
Let let's bring it back.
我们已经有了两个基本原则。
We have we have two solid principles.
要诚实，而且要让人觉得有人情味。
Be honest, and it should feel human.
你们两位还有什么想法，能帮我们总结一下，作为别人可以用的好准则吗？
Is there anything else from you two that you think we can wrap this up that's, like, a really good guideline others can use?
我觉得还有一点是关于模仿和借鉴的。
I think there's something around copying and imitation.
对。
Yes.
对，参考别人的作品，那肯定很重要。
Yes, references, for sure.
对，就是拿别人的东西当参考。
Yes, using something as a reference.
我觉得这是你应该遵守的，不管你是不是用AI，都不应该那么明显地模仿别人。
Think And this is something that you should abide by, in my opinion, whether you're using AI or not, you should not, imitate someone so clearly.
就像那句话说的，要像艺术家一样去偷。
Like this needs it's like steal like an artist.
所以你得去借鉴和参考，但你借鉴来的东西得转化到一定程度，让原作者看到你的作品也能接受。
So you have to take inspiration and reference, but where you take that inspiration has to go somewhere so far that that artist who originated that piece of work would be okay with what you've done with it.
这当然是很主观的。
And that's all very subjective.
不是，我是说，Austin Kleon来我们节目的时候也说过，YouTube社区里抄袭现象很严重，对吧？
No, but I mean, Austin Kleon said it to us when he came on the show, when we were asking, there's like a large amount of plagiarism in the YouTube community, right?
大家就是直接复制粘贴。
People just copy and paste stuff.
比如别人的缩略图、视频创意，全都是这样，对吧？
Like people's thumbnails, video ideas, all It's like of just constant, right?
这其实就是互联网的文化，但AI的出现加速了这种现象。
Like that is just the culture of the internet, but it has become AI has created an accelerant to that.
我觉得这应该是一条创作者守则，适用于所有场合。
And I think that is a rule, like a creator rule that should exist in every context.
就像一条艺术准则。
It's like an artistic rule.
就是，别抄袭我的东西。
Like, don't copy my stuff.
舞蹈界的规矩。
Dance creds.
确实。
For real.
对。
Yeah.
不，但你说得对。
No, but you're right.
你说得对。
You're right.
是的。
Yeah.
比如，我百分百会从别人那里汲取灵感。
Like, I take inspiration, 100%.
我们会参考一些作品，然后看看能从里面学到什么。
We take inspiration, we look at references, and we try and see where we can get
然后我们再创作出一些东西，
from And then we make something that
我们自己也
we are
真心觉得兴奋的东西。
honestly excited about.
对。
Yeah.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。