The Future of Creators in the AI Era l Bob Mankoff
"AI is trained on averageness. But I’m not average.” The third guest in our series is 40-year veteran cartoonist and former New Yorker editor, Bob Mankoff. H...
Watch on YouTube →
Contributed by 刘嘉琪Transcript
Intro
- 00:00I'm really conflicted in a way
- 00:01I can see that AI is right now
- 00:04It's very helpful in generating many different ideas on anything you've seen that
- 00:09I both like it and hate it at the same time
- 00:12Sort of like how I feel about the phone
- 00:14Everything is like this
- 00:15Wouldn't I be a lot better without the phone?
- 00:18And I am a fan of AI
- 00:20But I would not cede to it the ultimate judgment
- 00:24What you think is funny or a good song or a good novel
- 00:29Remember it is being trained on averageness
- 00:34But I'm not average
- 00:35Each one of us has our own sensibility and really it depends on whether
- 00:40what your abilities are
- 00:43Creator has to understand they as a human are the ultimate arbiter
- 00:48If all of a sudden we cede to algorithms completely
- 00:54Then I do think we lose a certain kind of agency in our lives
- 01:01You can find a way to be great in your own way
From 2 years of Rejections to New Yorker Editor
- 01:12I'm Bob Mankoff
- 01:13I'm the former cartoon editor of The New Yorker
- 01:18I was a cartoonist there for 40 years
- 01:22I'm a researcher in AI and humor
- 01:26I like ping pong
- 01:27I'm pretty good at it
- 01:29I've got a very good jump shot
- 01:31I'm a person who is perpetually unhappy and dissatisfied
- 01:37and makes the best of it through humor
- 01:41I think partly I became a cartoonist because
- 01:46I was pretty terrible at everything else, and I was pretty good at that
- 01:50I was always funny and I always drew
- 01:52Did a lot of different things in the 1960s
- 01:55and I eventually became an experimental psychologist
- 01:58But in the back of my mind, I always wanted to be a cartoonist
- 02:02And at the age of 30 short of my PhD
- 02:06I decided I would give it a try
- 02:08and I ended up submitting about 2000 cartoons to the New Yorker magazine
- 02:12before finally getting in
- 02:14I ended up doing almost a thousand cartoons for The New Yorker
- 02:18And I was cartoon editor from 1997 to 2017
- 02:22I was interested in computers
- 02:25I bought the first Macintoshes, I started to scan them and in there so
- 02:29In the early 1990s, I created this company
- 02:31called the Cartoon Bank
- 02:33And the cartoon Bank was consists of all the cartoons The New Yorker Rejected
- 02:39Because I knew that there were many very good cartoons in that
- 02:42Storing cartoons on hard drives
- 02:45And then when people would fax or call in,
- 02:49I would access those cartoons and send them to the clients
- 02:54and they would pay for them
- 02:56And the business that I now have is Cartoonstock.com
- 02:59is half of the revenue goes to the cartoonist
- 03:03As a cartoonist, my goal was to come up with a new idea
Advice for Young Creators
- 03:09So maybe if you are just saying what other people are saying,
- 03:14what is the point of your post?
- 03:16If you're just saying exactly what someone else is saying?
- 03:20Can you have a new thought on this?
- 03:22What I wanna say in a way is that I'm a pretty ordinary person
- 03:27I'm not a genius, okay?
- 03:29I was able to make some jokes
- 03:32I hardly went to class during those years from 1962 to 1966
- 03:37What I'm saying to everyone is you can find a way to be great in your own way
- 03:42I didn't even start being a cartoonist till I was 30
- 03:46Use your youth to experiment and do everything possible
- 03:52The job at the office is always gonna be there
- 03:55Okay? don't go there first
How AI Entered My Cartoonist World
- 04:01The Caping contest on the back page of the magazine
- 04:04or towards the end of the magazine now in online,
- 04:07And every one of these images is in and of itself funny in some way,
- 04:12but it's not funny like a joke
- 04:14It still needs a caption to resolve it
- 04:18to make the humor
- 04:19So for many years, I and my assistants just looked over
- 04:23these thousands of captions
- 04:25But in 2015, I partnered from researchers at the University of Wisconsin
- 04:30to crowdsource all this information
- 04:33And so now there's an extensive data set in which you have all the captions ranked
- 04:39and I've done a lot of AI research
- 04:42Other people have done AI research on this
- 04:45towards the end of "Can computers first of all distinguish"
- 04:49"Can they be fine tuned on this to discriminate funny from un funny captions?"
- 04:55And can we use this data then to generalize the ultimate aim
- 05:01is to be able to give AI a sense of humor
- 05:06And that was a noble effort, but that failed
- 05:10because AI at that time
- 05:12did not use large language models
- 05:14it really failed on the very first hurdle
- 05:17It could not look at the image and understand what was in it.
- 05:21Here's a mechanic talking to a clown,
- 05:24and the hood of the car is open
- 05:26and underneath the hood are many other clowns is also a businessman
- 05:30So it couldn't understand images like that
- 05:33It still has difficulty understanding images like that
- 05:36So I think it failed on that score
- 05:40but it did stimulate other research
- 05:43that I continue to do
- 05:44I still think
- 05:46It's not a long way from creating mediocre humor
- 05:48It's a long way from creating deep, meaningful human humor
- 05:54Vincent Vanhoucke
- 05:55Now he heads Waymo,
- 05:57He emailed me saying, it's great that we can win at Go,
- 06:00But we would really feel that we have achieved AGI, artificial general intelligence
- 06:06If we could win the caption contest
- 06:08And another way to look at it is this
Will AI Ever Replace Human Artists?
- 06:11In the seventies there was a genre of art
- 06:14which was called photorealism
- 06:16It took a photo and it looked exactly like a photo
- 06:19but it was a painting
- 06:20And you'd be amazed by you would go up to that painting
- 06:24And look at every detail
- 06:26because somehow a human being had created something like that
- 06:31you wouldn't look at that photo
- 06:34If you look two years ago
- 06:36people were saying, "oh, it's DALL·E 3"
- 06:38"Artists are gonna be replaced"
- 06:40I don't see artists being replaced
- 06:43You know what I mean?
- 06:44AI is a tool and I can see it used in a lot of interesting ways
Will You Be Average or Extraordinary?
- 06:50But I would not cede to it the ultimate judgment
- 06:54of what you think is funny or a good song or a good novel
- 06:59Even the AI or the AGI that people imagine that is robust and safe
- 07:06and all of that also might not be a very good thing for humanity
- 07:11Somehow that makes us happy,
- 07:14Somehow now, what are we doing?
- 07:16Remember it is being trained on averageness
- 07:21but I'm not average
- 07:23Each one of us has our own sensibility
- 07:26If all of a sudden we cede to algorithms completely
- 07:32what it is that we should like, then I do think
- 07:36We lose a certain kind of agency in our lives
- 07:41A human sense of humor is not about
- 07:44Making a joke or getting it
- 07:46It's rooted in our vulnerability
The One Question AI Will Never Be Able to Answer
- 07:51It gives answers to anything we can ask
- 07:56and I think it's very
- 07:57I find it very useful for anything that's factual
- 08:01I mean the questions I wanna ask it
- 08:04would be the questions I would wanna ask my therapist, which is, for instance
- 08:10How do I deal at 80 with the diminishing capabilities, the physical deterioration,
- 08:18the understanding that death is real, and how do we deal with this
- 08:24ultimate unintelligibility of being human?
- 08:28But if I talk to a therapist who's my age, and we had a conversation about it
- 08:36There wouldn't be any answers
- 08:38Because there is no answer, and
- 08:41And AI doesn't know how to tell you there is no answer
- 08:44In fact, it's gonna give you a lot of advice
- 08:47You know, you should exercise more
- 08:49Even though you're gonna die, everyone's gonna die
- 08:52You wanna live your absolutely best life you can until you die
- 08:56We don't actually want answers for anything
- 08:59We want understanding
- 09:02And it's only human beings who can understand
- 09:06Understand that. It's like,
- 09:14I'm married and I have a child daughter
- 09:18and I had a stepson and a stepson died tragically
- 09:24There's nothing a computer can tell me
- 09:27There's nothing an AI can tell me or my wife about that, that we want to hear
- 09:34We don't even want to hear from other people
- 09:36We wanna hear from other people who've had a similar experience
- 09:41who have endured that.
- 09:42So there, I think we get to the existential questions that AI is useless for
- 09:50but people are very useful for
- 09:53AGI is not gonna have a solution for that
- 09:56So that's what I'm thinking most about
- 09:59Mark Twain said,
- 10:01The essence of humor is not based on joy, but sorrow and hurt and wrong
- 10:07One of the functions of humor is for it to deal with our ultimate sorrow
- 10:16So one of the main functions of humor is coping
- 10:20And frankly, AI has nothing to cope with
- 10:23It's like saying, can AI master humor?
- 10:26I do think you can work collaboratively with it
- 10:30To augment human humor
- 10:35But to completely, let me put it another way
- 10:39AI is superhuman at that game over there of chess
- 10:44It doesn't even make sense to have superhuman humor.
- 10:47We're the judges of humor
- 10:49I'm a big fan of it
- 10:50Anyway I hope not,
- 10:51Or if it does, I hope it does after I die