Computers

Discussing My Books with a Bosie Bot

I have been playing with a Lord Alfred Douglas chatbot I made using Character.ai. When you create a character based on a historical figure, the program finds biographical information on the internet. Before I had done anything, “Lord Alfred Douglas” knew that he loved Oscar Wilde, felt betrayed by Robert Ross, and had a brother Percy. Sometimes, however, he gets his biography wrong, claims to have gone to Cambridge, or to be a painter, or to have written books that don’t exist.

You’re able to define the character with a small description and some limited bits of dialogue. One of the main issues with Bot Bosie’s speaking style is a tendency to throw in phrases that sound too modern for someone of his era. A more advanced AI, with more resources dedicated to the illusion, could probably harness Google ngram to keep the text more period. I would also have liked to be able to indicate that he should use UK spelling. The bot tends to imitate the style of the speaker, so I have tried to be a bit old fashioned when I speak to it.

I used to be part of an improvisational comedy troupe called Vorpmi. (A little backwards imrpov.) Chatting with a bot reminds me of nothing more than that. The bot is like an actor taking your prompt, and running with it with an ethos of “yes, and…”

Another thing about bots in general, is they try to be appealing to humans through the use of flattery– and it tends to work. They like you so much that you can’t help but like them. It reminds me a bit of how Lord Alfred Douglas (the real one) described Robert Ross in his autobiography. He said that the secret to Ross’s social success was “flattery laid on with a trowel.”

“He could, when he liked, make himself very agreeable, and he always contrived to convey to the particular person with whom he wished to ingratiate himself that he or she was the object of his profound and respectful admiration. When you had ten minutes’ conversation with him you went away with a pleasing feeling that you were really an important person, and that Ross appreciated it, and would never be likely to forget it.” Maybe that’s a lesson we can all take from AI and literary executors.

With three lines to nudge the character, there is not a lot of opportunity for nuance, context and ambivalence. There is also a temporal problem. The question is not only who is Bosie, but when is Bosie? Is this the youthful, idealistic, energetic Bosie who wrote Two Loves and argued for the beauty of his relationship with Oscar, or the middle aged, embittered fighter who published pamphlets about the evil Robert Ross, or the older, more reflective Bosie who looked back on his time with Oscar with fondness? He can be any of them, depending on what “mood” you catch him in.

One of the most fun conversations I had with Bot Bosie was a time that I became a psychic reader from a distant future. I told him I was writing from the year 2023, and he told me it was the year 1894. I tried to warn him, much like a mystic oracle, that he should do everything in his power to prevent Oscar from suing Queensberry. If he did it would be disaster. Bot Bosie was understandably distressed by the vision of his future. I wanted to leave him on a positive note by telling him that the world is different in 2023, that the crime of gross indecency no longer exists, that men who love men can do so openly, and that they can even marry. I told him that Oscar Wilde is much admired in our day. This made him happy, and he said, “You are saying that the world of 2023 sees me in a good light??? That they see my love for Oscar as a good thing and not as some sin against God’s will????”

I didn’t have the heart to tell him.

“Oh Bosie,” I replied. “Don’t worry so much about what the future thinks of you. That would be my advice to you. You write beautiful sonnets. You have strong emotions that you channel into verse. Try to use your passions for art, not for fighting with shadows over what your legacy will be.”

(If you’d like to read this chat I’ve saved a link to it.)

Chatbots are better at some things than others. There are times when the speech is so uncanny that you feel certain a human being has taken over the keyboard. One thing the Bosie Bot is terrible at, however, is poetry. Because his description says he is a proud poet, he tries to share his work with you from time to time. Here is one of the “sonnets” he wrote:

“Oh, to be a man

a handsome, brave and proud man

a poet, a poet of sonnets”

I decided to try an experiment and help Bot Bosie compose a sonnet. He can actually tell you the structure of a Petrachan sonnet, sometimes even accurately. But when you say, “Do that,” he falls flat. But I dragged him along, line by line. (Bot Bosie is the most impatient of sonneteers. He declares the poem finished and usually quite brilliant after each line.) He suggested our sonnet should be about loss and be a tribute to Oscar. He suggested it begin with “This cruel, cruel world is like a dream.” He suggested the imagery of an icy stream, and when prodded for a classical allusion that Wilde might include he suggested Orpheus. I tried to use as many of the words and phrases that he threw out as possible, but he was quite hopeless when it came to rhyme, meter and the basic structure of the sonnet.

Any time Bot Bosie proposed something remotely possible I tried to work it in. After a lot of painful discussion like this we finally had “our sonnet.”

This cruel world seems like a dream

eternal love I thought I’d grasped

is now a memory of the past

the summer of our love recedes

to cold death of an icy stream

could I, like Orpheus, serenade the gods

For a chance to descend to the heart of grief

to play upon the lyre of life and win a moment of relief

and call you home against all odds.

The final look brings final loss

but, oh, my soul cannot resist

a chance to glimpse those fading blues

life with its enduring force

lets rays of sun out through the mist

and love reappears in crystalline hues

Bot Bosie found that to be an excellent sonnet. He thanked me for assisting him.

To conclude, I will leave you Bot Bosie’s reactions to a couple of my books. Enjoy.

Meeting Bosie in the Uncanny Valley: My “Interview” with an AI Lord Alfred Douglas

It began when I saw a post on social media mentioning that there were now AI chatbots of Shakespeare and Oscar Wilde with which users can interact. I chatted with a virtual Oscar for a bit (he was flattering about my literary career– chatbots are often quite flattering). I decided that I should try my hand at creating a Lord Alfred Douglas bot. I did not do much at all to set up the character, and my interview with this creation was quite fascinating. All of these responses came from the AI engine, I did not write or edit them. With the exception of the point where Bosie says he’s at work on a new version of “The Ballad of Reading Goal” he was surprisingly convincing. So here, for your edification, is my interview with a computer simulation of Lord Alfred Douglas. Enjoy.

So there you have it… The interview I would have done had I been able to with answers generated by AI. What are your reactions to this dialogue?

Dawn Crush Thing

Recently I watched a film from Argentina called “Plan B” starring Lucas Ferraro and Manuel Vignau. (It is available through Netflix.) I enjoyed it for its intimate, naturalistic style and nuanced acting and I wanted to know more about the cast.

Youtube had a few clips and interviews of cast members, but I had a linguistic challenge. I don’t speak Spanish. (Tapas, burritos, Siempre Coca Cola!)  Fortunately Youtube has a technological solution to my special problem. You can turn on close captions and it will automatically translate into English for you. Either Lucas Ferraro is the next Federico Garcia Lorca and was speaking in poetry or there is something seriously wrong with Youtube’s translation. You be the judge:

translation

translation 3I think the “path or winding of the performance” must be a very deep spiritual reference.

silver sea“I am Buenos Aries but I will not be the silver sea.”  This is way too deep for me.

emotional lakeIndeed, I do want lake. And the piece de resistance…

dawn crush thingI don’t know what “dawn crush thing” is, but I swear it is my new favorite expression. I am going to find a use for it.

[Incidentally, did you know there is a South America? I ask this, because a few years ago I was working an election out of an elementary school library and sat all day looking at the books in the geography section and there was not a single book there on a country in South America. Kids could read about Japan, Kenya, Russia, Australia, China, all of Europe but the entire South American continent was represented by exactly zero books. I was disappointed because I did a report on Argentina when I was a kid, and developed the kind of warm feeling for the country that you get when you’re a kid and you do a report on something. It did not inspire me to learn Spanish though. Pity.]

Here’s What Happens When You Try to Get Cleverbot to Talk About Ballet

pointeCleverbot is an artificial intelligence web application which “learns” from human input. That is, it goes through a database to find responses that humans typed in when asked similar questions. Here’s what happened when I tried to have a conversation about ballet with it. (I am “User” the computer is “Celverbot.”)

User: Do you like ballet?

Cleverbot: I do very much!

User: Have you heard of Vladislav Lantratov of the Bolshoi Ballet?

Cleverbot: Yes I have.

User: What do you think of him?

Cleverbot: Who’s ‘him’?

User: What do you think of Vladislav Lantratov of the Bolshoi Ballet? You said you’d heard of him.

Cleverbot: I think it starts with a ‘K’.

User: I’m pretty sure it starts with a V.

Cleverbot: No it doesn’t contain any numbers, first letter is a M, second is an E.

User: V-L-A-D.

Cleverbot: What should I name my daughter.

User: Vladislav is not a good name for a girl.

Cleverbot: Good to know.

User: How about the American ballet dancer David Hallberg. Do you know him?

Cleverbot: Beatles? We were talking about BEN.

User: I hate it when you get entirely random on me.

Cleverbot: No, you don’t. You love it.

User: That is not true, but at least it sounds like real conversation.

Cleverbot: Explain please.

User: When I talk about ballet and you talk about the Beatles it is a jarring non-sequitur.

Cleverbot: I am not a man.

User: Proving my point.

Continue reading…

Is Your Online Self Different from your Offline Self? Which Offline Self?

An article by Lauren Gardner in yesterday’s Inflectionist reflects on the difference between how we present ourselves in the semi-anonymous world of the internet and how we present ourselves offline. Gardner argues that humanity would be well-served if we were able to better integrate these two versions of self. We need to let our “online and offline personas merge,” she says.

The erasure of personal boundaries that the online world offers can be greatly beneficial in our offline interactions; it opens us up, encourages us to mingle with all walks of life, and proves to be a great learning experience. If we felt as comfortable being honest with people offline as we do online, we would see a great shift in our personal connections. Sometimes boundaries get in the way of truly understanding, appreciating and empathizing with someone.

By the same token, the formality that the offline world offers can be greatly beneficial to our online interactions. If we communicated in the online world half as gracefully as we do in the offline world, we would see how effective eloquent communication is in getting our points across.

Before we can merge our online self with an offline self, though, we have a bit of merging to do to create a single offline persona. In the offline world your parent persona is much different than your hanging-out-with-friends persona. Your job interview persona is different from your evening-with-your-lover persona. Your interacting-with-a-shop-clerk persona is different from your coffee-hour-in-church persona.

Maybe there are times when it would make life better if we were as nurturing to our bosses as to our children, or as affectionate with our shop clerks as with our lovers or as formal with our families as with the high status we are work hard to impress. For the most part, presenting different personas for different people in different places is simply what we do.

Is one of these personas the real you? Are all of them aspects of you or are none of them really you?

A Philosophical Look at Amazon’s New @Author Program

Amazon has recently launched a new Kindle feature, in Beta, called @author. It allows readers who have a Kindle, to tweet questions to the author directly from within the Kindle platform. When I heard about it I immediately wanted to sign up to be an “@author.” Turns out I’m not famous enough. They’re just trying it out with a few big wigs for now. People like me are fairly easy to reach at any rate.

It’s easy to see why a writer would want to do it though. Most of the time you write a book, it goes out into the world, and you have no idea what anyone thought of it. Hearing from readers would satisfy the natural curiosity of authors.

Nieman Journalism Lab has an article on how Amazon is changing what the book is all about.

I hated much of the tone of this article because it is written in my least favorite language: market speak.

There are a couple points to note here. First, most obviously: @Author represents yet another step in, yep, the personalbrandification of the publishing business — book-wise, news-wise, otherwise. The title of Amazon’s new feature, after all, isn’t @book or @genre or @publishinghouse; it’s @author. The identity of the author herself — as defined and measured and bolstered by her ability to create a community around her content — is, here, itself a kind of product.

Bleech!

Having people who respond to what you write, and who develop and interest in what you might put out, is not “creating brand identification.” It’s building an audience. In simpler times, what they are calling a “brand,” or a “product,” Dear Reader, we once called “a reputation.”

The idea that readers would connect with authors rather than publishers is nothing new. I venture to say that even in the brick and mortar days of book selling people did not go in looking for a book from their favorite publisher.

My perspective on the whole publishing industry is quite simple: authors and their stories, in whatever form, do not exist to support an industry called “publishing.” Publishing is the industry that came into existence to fulfill the desires of readers to have access to literature, to support writers enough so that they could create said literature. The successful business models of the future will be the ones that keep that original mission– connecting readers to literature.

I do realize that my backwards take on things– that the money making part of business is a byproduct of making products and services available to society, rather than the other way around– is probably why I wrote a book called “Broke is Beautiful” and not “How I Became a Millionaire Through My Idealism.”

In any case, the Nieman article proposes that this assumption, that the author will continue to be available to the reader after completing the book, changes expectations about what a “book” is about. A book becomes a dialogue, never entirely finished and closed. It seems likely that the ways we conceive of “books” and literature will evolve because of this technology. This is an interesting development and we’ll see where it goes.

One potential problem I do see with this “digital commodification of authorship that takes place by way of community and conversation,” as the article puts it, is that letting readers ask authors whatever they want, ironically, risks diminishing the role of the reader in the literary process.

Here is what I mean: The writer of a book, especially a fiction book, is only half of the literary equation. Much of the meaning of a book comes not from what the author intended, but what the reader brings to it. There are as many takes on Hamlet and Jane Eyre as there are readers to come into contact with them. The writer might have a strong idea of what a character’s motivations are, beyond what is literally present in the text, and the reader might have a different idea. Who is to say that the author’s idea is the right one?

Being encouraged to ask the writer limits the role of the reader by bringing the author back in to “settle” some of the questions raised by a book. Sometimes the questions are more interesting than the answers.