This is the twenty-third in a series of interviews with explorers, thinkers, artists, activists and other luminaries around the world, people whose life’s work resonates with our founding principles.

这是对世界各地的探索者，思想家，艺术家，活动家和其他名人的系列访谈中的十二分之三，这些人的生活与我们的创立原则相呼应。

Our friend Christine McLeavey Payne is a Member of the Technical Staff at OpenAI, where she recently published MuseNet and Jukebox. Christine studied physics at Princeton, and neuroscience and medicine at Stanford. She’s a trained classical pianist and concert musician.

我们的朋友恭McLeavey佩恩是在技术人员的会员OpenAI ，她最近出版的MuseNet和点唱机。 Christine在普林斯顿大学学习物理学，在斯坦福大学学习神经科学和医学。 她是一位训练有素的古典钢琴家和音乐会音乐家。

Not one to shirk from healthy skepticism and speculation, Christine and I chatted about the present and future of AI and all of its myriad impacts on society and the workplace.

我和克里斯汀·克里斯蒂安(Christine)和我(Christine)谈不上健康的怀疑和猜测，谈论了人工智能的现在和未来以及它对社会和工作场所的诸多影响。

GP: You’ve spent the last few years using AI algorithms to generate novel music. You’re also trained as a classical pianist. Have algorithms learned to develop their own style? How might a real life human musician interact with a composing algorithm?

GP：过去几年，您一直在使用AI算法来产生新颖的音乐。 您还会接受古典钢琴家的训练。 算法是否学会了发展自己的风格？ 现实生活中的人类音乐家将如何与作曲算法互动？

CMP: Models like MuseNet, Jukebox, and other AI music algorithms are still very much the early steps — they’re like students at conservatory who are learning how to imitate the great musicians of the past.

CMP：MuseNet，Jukebox和其他AI音乐算法之类的模型仍是非常早期的步骤-就像音乐学院的学生一样，他们正在学习如何模仿过去的伟大音乐家。

I love when MuseNet suggests a new and surprising way a Mozart sonata could have gone, or when Jukebox generates a sample that sounds like Ella Fitzgerald’s voice, but to me the really exciting moment in music will be when future models can generate totally new sounds and styles that are also meaningful to us musically and emotionally. I’m looking forward to the point when I’m truly shocked and moved hearing a great idea I’d never even imagined before. We’ve seen superhuman chess, or superhuman Go, and sometimes I wonder if there’s conceivably such a thing as superhuman art.

我很喜欢MuseNet提出的莫扎特奏鸣曲可能以新的令人惊讶的方式出现，或者当Jukebox产生听起来像Ella Fitzgerald的声音的样本时，但是对我来说，音乐中真正令人兴奋的时刻将是将来的模型可以产生全新的声音并在音乐和情感上对我们也有意义的风格。 我期待着当我真正震惊和感动的时候听到一个我从未想过的好主意。 我们已经看过超人国际象棋或超人围棋了，有时我想知道是否存在像超人艺术这样的东西。

I’ve worked with several composers who have tried working with MuseNet, and my favorite setup at the moment is to train the model on that composer’s past pieces. Composers describe the surprise of hearing the model imitate their own voice, saying it feels like something they might have written, but nothing they’d actually done in the past. Then the composer can write the start of a piece, and MuseNet will suggest many different ways to continue, and there can be a sort of back and forth dialog. These models are very good at generating many different options quickly, and some are good and some terrible. In some ways they’re the perfect brainstorming partner.

我曾与几位尝试使用MuseNet的作曲家合作，而目前我最喜欢的设置是在该作曲家的过去作品上训练模型。 作曲家形容听到模型模仿自己的声音感到惊讶，说这听起来像是他们可能已经写过的东西，但过去却没有做过。 然后，作曲家可以编写乐曲的开头，MuseNet将建议许多不同的继续方式，并且可以进行某种来回对话。 这些模型非常擅长快速生成许多不同的选项，有些则很好，有些却很糟糕。 在某些方面，他们是完美的头脑风暴伙伴。

演示地址

GP: Are machines now capable of genuine creativity? How long before we popularize the masterpieces of an algorithm? What does this mean for creative attribution? Would you, as the creator of the algorithm, be credited with composing a sublime piece of music?

GP：机器现在具有真正的创造能力吗？ 我们普及算法杰作需要多长时间？ 这对创意归因意味着什么？ 作为算法的创建者，您会因创作出色的音乐而受到赞誉吗？

CMP: I wouldn’t be at all surprised if an AI song is popular in the next few years, although that’s more because the music industry itself has a formulaic approach to writing the really big pop song hits. I think it’ll be further off until we hear an AI masterpiece that’s totally new and creative and surprising, and not just a blend of existing human styles.

CMP： 如果AI歌曲在接下来的几年中流行，我一点也不感到惊讶，尽管那更多是因为音乐行业本身拥有一种公式化的方法来编写非常流行的流行歌曲。 我认为这将是遥不可及的，直到我们听到一个AI杰作，它是全新的，富有创造力的和令人惊讶的，而不仅仅是现有人类风格的融合。

The attribution part is tricky — it’s important we don’t give full copyright protection to AI songs. At that point, what’s to stop someone from setting a machine going and generating hundreds of thousands of songs, and then claiming copyright on the whole collection? Still, the creator of the algorithm deserves some credit in the composition of the music.

归因部分是棘手的-重要的是，我们不对AI歌曲提供完整的版权保护。 到那时，什么阻止了某人设置机器运行并生成成千上万首歌曲，然后要求对整个唱片集主张版权？ 尽管如此，该算法的创建者在音乐创作方面还是值得赞扬的。

At OpenAI, music research is just part of our overall mission, so we’ve released the MuseNet and Jukebox models freely for use, just asking that people cite that they used our work. However, if an independent programmer works for years creating an algorithm that generates the first big hit song, we’ll need to figure out some way that they get fairly compensated for that.

在OpenAI中，音乐研究只是我们总体使命的一部分，因此我们免费发布了MuseNet和Jukebox模型供使用，只是要求人们引用他们使用了我们的作品。 但是，如果一个独立的程序员工作了多年，创建的算法可以生成第一首大热门歌曲，那么我们需要找出某种方式使他们得到相应的补偿。

演示地址

GP: Has our current predicament accelerated the development of teaching AI’s? Will most teaching happen in dialogue with algorithms in the future? Who gets to decide what the AI’s are allowed to teach? Why does this matter?

GP：我们当前的困境是否加速了AI教学的发展？ 将来，大多数教学会与算法对话进行吗？ 谁来决定允许AI教什么？ 为什么这么重要？

CMP: I’m really excited about the potential for educational AI — I think it could be one of the earlier ways AI can truly benefit society at large.

CMP：我对教育性AI的潜力感到非常兴奋-我认为这可能是AI真正使整个社会受益的早期方式之一。

During the pandemic, I’ve spent a lot more time sitting and teaching my own kids. We’re lucky in to be in a great school district with fantastic teachers, but even still I’ve seen how much more quickly my kids are progressing with the constant one-on-one attention. We can follow whatever topics my kids are naturally excited about (right now we’re doing a lot of Minecraft math!).

在大流行期间，我花了很多时间坐下来教自己的孩子。 我们很幸运能在一个很棒的学区拥有出色的老师，但是即使如此，我仍然看到我的孩子在持续不断的一对一关注下进步得更快。 我们可以跟随孩子自然兴奋的任何主题(现在，我们正在做很多Minecraft数学！)。

I also feel acutely aware that my husband and I are lucky to have jobs where we can work from home during Covid, and can spend this much time with the kids (and of course there have been plenty of times where we’re in work crunches and haven’t been able to be so focused!). I find myself frequently wondering if the way to scale this individualized attention could be through AI.

我也很敏锐地意识到，我和丈夫很幸运能够在Covid期间有在家工作的机会，并且可以花很多时间陪伴孩子们(当然，很多时候我们都在做工作)而且还无法如此集中精力！)。 我发现自己经常想知道是否可以通过AI来扩大这种个性化关注的方式。

Imagine if we all had personalized AI tutors — encouraging, patient and friendly, and exactly suited to our own skills and interests. Right now, if I want to learn a new skill, I hunt for a blog, or a textbook, or an online course. There’s usually something close, but it’s never exactly tailored to how much I already know.

想象一下，如果我们所有人都有个性化的AI导师-鼓励，耐心和友好，并且完全适合我们自己的技能和兴趣。 现在，如果我想学习一种新技能，我会去寻找博客，教科书或在线课程。 通常有一些接近的地方，但从来没有完全根据我已经知道的程度量身定制。

But what if I could talk directly with this tutor, and it could provide me with precisely the background I was missing. It could pick up on when I’m still confused, or when I’ve learned something easily and want to jump ahead. It could notice when I have energy or when I’m distracted — let me know I should take a break to stretch, or turn things into a silly educational video game. Imagine that all kids had access to this sort of hyper-personalized, fun, and supportive education. (Obviously we still have the huge problem that many kids don’t have access to laptops and wifi, but I do feel hopeful this will be an increasingly solvable problem.)

但是，如果我可以直接与这位导师交谈，并且可以为我提供我所缺少的背景，那该怎么办？ 当我仍然感到困惑时，或者当我轻松地学到一些东西并想向前迈进时，它可能会开始上升。 它可以注意到我什么时候有精力或什么时候分心–让我知道我应该休息一下，或者把事情变成一个愚蠢的教育视频游戏。 想象一下，所有孩子都可以享受这种超个性化，有趣和支持性的教育。 (显然，我们仍然面临很多孩子无法使用笔记本电脑和wifi的巨大问题，但我确实希望这将是一个越来越可解决的问题。)

Of course, I then wonder who will decide what curriculum the AI teaches. Already as a society we argue about what school standards should be — if we include creationism, or anti-vaccination arguments, if we focus on our own country’s history, or adopt a global view, how we look at balancing capitalism and environmentalism. In some ways, our safety net is that education is still spread out — some kids will grow up being taught that abortion is never acceptable, and others will grow up learning about flying spaghetti monsters, and this helps us to preserve our essential diversity of thought.

当然，然后我想知道谁来决定AI教授什么课程。 作为一个社会，我们一直在争论应该采用什么样的学校标准—如果我们包括创造论或反疫苗论据，如果我们关注本国的历史，或者采取全球观点，我们将如何平衡资本主义和环境保护主义。 在某些方面，我们的安全网是教育仍在传播–一些孩子将长大，被告知堕胎是永远不能接受的，而另一些孩子将长大，学习飞行意大利面的怪物，这有助于我们保持思想的基本多样性。

If twenty years from now, we have an AI tutoring system in place of classrooms, who will decide what the AI teaches? Will it be the 50 or 100 members of some startup company? Will it be Google or Facebook? Even if great teaching AI is still a little ways off, I think these are topics we should all start discussing now.

如果从现在开始二十年，我们将拥有一个AI辅导系统来代替教室，谁来决定AI教什么？ 它会是某个创业公司的50或100名成员吗？ 是Google还是Facebook？ 即使出色的AI教学距离尚有一段距离，我认为这些都是我们现在应该开始讨论的主题。

GP: In 1930, economist John Maynard Keynes forecast that his grandchildren would enjoy a 15 hour work week. Obviously this hasn’t occurred. What is our happiest future in terms of automation? Will we finally arrive at something like the 15 hour work week?

GP：1930年，经济学家约翰·梅纳德·凯恩斯( John Maynard Keynes)预测，他的孙子每周工作15小时。 显然这还没有发生。 就自动化而言，我们最快乐的未来是什么？ 我们最终会达到每周15小时的工作时间吗？

CMP: I love this idea! I’d actually never heard that forecast before. Honestly I do notice so frequently in my own work that I’m way more productive in a week if I focus on getting 3 *amazing* hours of work done each day, rather than focusing on long hours. I would love if AI helps bring this on — both by helping take away the mundane frustrating work, and by helping us to be focused and at our peak.

CMP：我喜欢这个主意！ 我实际上从来没有听说过那个预测。 老实说，我确实在自己的工作中经常注意到，如果我专注于每天完成3个*惊人*小时的工作，而不是专注于长时间的工作，那么一周内我的工作效率就会更高。 我希望AI能够帮助实现这一目标-通过帮助消除平凡而令人沮丧的工作，并帮助我们集中精力并达到巅峰。

Still I think part of this shift will have to come in societal expectations — especially in the US, we have such a sense that we’re supposed to be busy all the time. We also have a lot to fix economically — for example, I’m not sure how we’d ever convince companies to give full health care and benefits for 15 hour work weeks, so it’s likely something we’d have to implement at a societal level.

我仍然认为，这种转变的一部分必须来自社会的期望-特别是在美国，我们有一种感觉，我们应该一直在忙。 我们在经济上还有很多需要解决的问题-例如，我不确定我们如何说服公司在15小时的工作周内提供全面的医疗保健和福利，因此我们很可能必须在社会上实施这些措施水平。

Still, this sounds like an amazing goal to me. I don’t worry that people will be totally unmotivated if they no longer have the financial strain to work incredibly long hours. People are also so motivated by wanting to feel useful and respected.

尽管如此，这对我来说仍然是一个了不起的目标。 我不用担心，如果人们再也没有财政压力来长时间工作，他们将完全没有动力。 人们也因为想要感到有用和受到尊重而受到激励。

GP: You and I met through Renaissance Weekend, I believe through our mutual friend and the Executive Director, Ali Gelles. Why do you and Steve keep going back? What is so special about this community?

GP：您和我是在Renaissance Weekend期间相识的，我相信是通过我们共同的朋友和执行董事Ali Gelles。 你和史蒂夫为什么继续往回走？ 这个社区有什么特别之处？

CMP: To me the most striking thing about RW is the sense of service. The sense that it’s a tremendous privilege to choose a career out of interest rather than necessity, and with that comes the responsibility to choose work that is meaningful.

CMP：对我而言，RW最令人震惊的是服务意识。 选择职业是出于兴趣而不是必要，这是一种极大的特权，随之而来的是选择有意义的工作的责任。

I always end up leaving RW gatherings worrying about some new world problem I’d never even heard of before, but also feeling reassured that there are very good and very smart people working and caring deeply about these problems.

我总是最终离开RW大会，担心一些以前从未听说过的新世界问题，但也让我感到放心的是，有非常优秀，非常聪明的人正在努力和关心这些问题。

GP: Have our current apocalyptic circumstances dampened your enthusiasms? Are there any obvious silver linings in all of the change we’ve experienced since March? How are our current struggles playing out amongst families and single friends?

GP：我们目前的世界末日情况是否挫败了您的热情？ 自三月以来我们经历的所有变化中是否都有明显的一线希望？ 我们目前在家庭和单身朋友中的斗争如何进行？

CMP: This has been an unbelievable year! More than usual, it feels like we’re divided by each facing very opposite kinds of problems.

CMP：这真是令人难以置信的一年！ 比平时更多，感觉就像我们被每个面临的非常相反的问题所分开。

At work, there’s a clear distinction between the people who are single and those with families. The single people are often struggling with feeling lonely, but have often been wildly productive at work. Whereas the people with kids don’t face the loneliness in the same kind of way, but are constantly swamped and exhausted. I feel for the teachers who understandably don’t feel safe returning to in person teaching, and also the parents who have work that can’t be done from home and who need a safe place for their kids to learn. I worry about my musician friends who struggle instead with the uncertainty of when or if work will ever return.

在工作中，单身人士和有家人的人之间有明显的区别。 单身的人常常在孤独中挣扎，但在工作中却经常表现出疯狂的生产力。 有孩子的人不会以同样的方式面对孤独，而是不断地被淹没和疲惫。 我感到可以理解的是，对于那些无法接受亲自教学的老师，以及那些无法在家完成工作并且需要一个安全的地方供孩子学习的父母，我感到可以接受。 我担心我的音乐家朋友会为何时或是否能够返回工作的不确定性而苦苦挣扎。

The big silver lining for me has of course been spending this much time with my kids and watching them grow very close as brothers. One of my other favorites has been discovering small Zoom classes around the world about various topics. Back in the spring I took a few Italian cooking classes — 4–5 different families all around the world, all stuck home in the pandemic, but sharing a few hours cooking delicious pasta together and talking. I used to be a completely terrible cook, so I’ve really appreciated that these months have forced me to learn to be more confident in the kitchen. More recently I’ve been practicing French and Spanish by finding online meetup groups, again with people around the world who want to talk. My kids are taking Capoeira Zoom classes from a teacher several time zones away.

当然，对我来说最大的一线希望就是花很多时间陪伴我的孩子，并看着他们与兄弟们变得非常亲密。 我的其他最爱之一是在世界各地发现有关各种主题的小型Zoom课程。 早在春天，我参加了一些意大利烹饪班–全世界4-5个不同的家庭，全都住在大流行病的家中，但是共享几个小时一起烹饪美味的意大利面并聊天。 我曾经是一个完全糟糕的厨师，所以我真的很感激这些月迫使我学会对厨房更加自信。 最近，我通过寻找在线聚会小组来练习法语和西班牙语，再次与世界各地想聊天的人交谈。 我的孩子们正在几个时区之外的老师那里上Capoeira Zoom课程。

I’m so looking forward to gathering again in person with friends, but I do hope post-pandemic we can keep this sense that long distances and friendships can feel as close as nearby ones. I love the new tradition of staying connected to old friends from around the world through Zoom meals, morning workouts, and happy hours, and I really hope to continue that.

我非常期待与朋友再次聚会，但我希望大流行后我们能保持这种感觉，即长距离和友谊可以和附近的朋友一样亲密。 我喜欢通过Zoom进餐，晨练和快乐时光与来自世界各地的老朋友保持联系的新传统，我真的希望继续下去。