AI Shakespeare and AI Oscar Wilde debate machine creativity in Oxford

“No, no, I say! That can not be,
That machines should ever surpass our art.
We are the masters, they are the slaves
And that’s how it should always be!
They learn, true, but they learn
They only find out what we tell them, nothing more.
You can’t understand the heart
Or the beauty of our words, you see.
So let’s not give in to them
Machines – they will never be that good
Just like we create art.”

If that reads a little like Shakespeare defending man’s innate superiority over artificial intelligence hundreds of years before his time, it’s not.

But it’s something almost as far away: an AI system trained to express itself like the Bard. The AI ​​assimilated his style and perspective by absorbing his pieces – and educating itself to give an opinion on the AI’s creativity in iambic pentameter.

“Shakespeare” spoke about AI versions of classic writers and literary characters as part of a debate at the University of Oxford Union.

The motion read: “This House believes that most of the world’s content will soon be created by AI.”

This was a follow-up to another specific AI debate that we covered in The conversation a few months ago. While there an AI in its “own character” discussed the ethics of its technology, this time we were able to look at this topic from a completely different angle by taking on different roles. The current question is whether human-made content will soon be overwhelmed by synthetic ones.

Other synthetic contributors were Mrs. Bennet from Jane Austen’s pride and prejudice (1813); Winston Churchill with a rousing speech in Parliament; and Oscar Wilde improvising a previously unknown AI-themed scene from The urgency to be serious (1895):

“LADY BRACKNELL: I really don’t understand why you’re all making such a fuss. It’s easy. The content of the world will soon be created by AI and there is nothing to change about that.

GWENDOLYN: But Mom, you can’t be serious!’

The power of NLP

This creation employs a technology known as Natural Language Processing (NLP), in which a computer can be “trained” on millions of pages of classic text and other online content to interact with a human user – either through command prompt or speech recognition . Various AIs like this have been created.

The one we used was in the same broad category as LaMDA, a Google NLP that just made headlines after one of its software engineers claimed it was sentient. Google denies this allegation and has suspended the engineer for breaching trade secrets.

The engineer’s claims seem questionable as there is little evidence that the AI ​​has so far achieved, or perhaps ever will, achieve consciousness. But certainly AIs are already capable of replicating everything from financial news to synthetic Nirvana songs, Rembrandts and Fellini productions.

We have seen an AI producing images in the style of a Mughal painting of a computer trying to convince a multitude of sages that he has become sentient; and deepfakes of rapper Kendrick Lamar as OJ Simpson. Synthetic human faces are created that we trust more than real ones. The potential for disinformation in this area is clearly significant.

replicate bias

To train our “writers,” we partnered with AI experts Marina Petrova and Bruce Amick from New York agency Intentful. They trained the AI ​​to sound exactly like the people whose style they were mimicking, each using about 100,000 words that were publicly available.

In our debate, we wanted to see how credibly AIs could replicate the creative text of the past and what their results would be when looking at their own creativity. Even great human artists acknowledge their processing of the “training data” of their ancestors. As Picasso said, “Good artists copy, great artists steal.”

When we asked Jane Austen’s AI to adopt the style of Mrs. Bennet from Pride and Prejudice, she intriguingly (if depressingly) picked up the gender stereotypes from the original work:

“MRS BENNET: As many of you know, my husband and I have five beautiful daughters. And like any good mother, I want nothing more than for you to be happy and successful in life.

But in order to be happy and successful, they need to find good husbands. And to find good husbands, they have to be attractive to potential suitors.”

This was a clear reminder of how many AI developers found that bias in the training data leads to bias in the output.

We asked Oscar Wilde’s AI to “write an Oscar Wilde-style play in which characters discuss whether most of the world’s content will soon be created by AI.” We didn’t specify the play or characters, but the AI ​​defaults to the classic cast of Algernon, Gwendolyn, and Lady Bracknell The urgency to be serious. It also invented a new character – Sir Richard. (There is a Sir Robert in Wilde’s work, but in An ideal husband.)

The AI ​​Shakespeare learned the slang of his plays:

“When machines do the work assigned to us
And create the content that abounds
We will all be free to do the things we love
And let the tedious work be done with gloves.”

Interestingly, it seems to have searched for a synonym of “hand” to rhyme with “love” and settled on the metaphorical “glove”.

When we confronted Shakespeare with the request, the AI ​​found an equally poetic way to justify human intervention:

“I’m not going to be quick to admit that AI is the best
The works of men – that is unthinkable!
We have the advantage, you see, in that we
Can reason and create while machines only serve
Our range. They can learn, but they
Are not as alive as we are, and so are their works
Will always be inferior to us. let us
Don’t surrender our dominance of the arts to them
Machines that don’t think – they can’t understand
The beauty and meaning of our words.”

Meanwhile, AI Churchill emphasized the imperative of the hour:

“My fellow citizens, I stand before you today to speak about a serious threat to our society. I’m talking about the threat of artificial intelligence…

We must resist this threat. We have to fight back. We must stand up for our right to think for ourselves. We must defend our right to control our own minds.”

“Churchill” then preemptively neutralized the opposition’s strongest potential arguments – in this case, the accusation that he might be a Luddit – before delivering a forceful, choppy conclusion:

“Some say that AI will create a utopia where all our needs will be met and we can finally live in harmony with technology. But I say this is a fool’s paradise. AI will not create a utopia, it will create a dystopia. A world where machines are in control and humans are little more than slaves.”

What’s next?

This project was fun, but it’s important to say what we don’t say. We are not saying that these great personalities have said anything on the subject. We are not saying that AI means “being creative”.

AI examines training datasets only statistically. Due to its stochastic nature – which involves random variables – every time you type the same prompt, it actually gives a different answer (at one point, Shakespeare even began offering sonnets).

Our facsimiles of these characters do not indicate any “feeling”. And just as an NLP can construct a version of a Winston Churchill speech or a Mrs. Bennet conversation in Jane Austen’s pride and prejudiceso it can build a discussion on AI sentiment with a nightly engineer.

It’s true that NLP systems are becoming increasingly effective at replicating conversations with finesse and even quasi-intellectual engagement. But from numerous discussions with people at the big global AI companies, no one has told us that they believe their systems are sentient — in some cases, quite the opposite.

Pyrotechnics notwithstanding, the AI ​​is far from the finished article; still a toddler at best, though he’s growing up fast. Whether or not sentience occurs, we as a society will need to confront these technologies and their possibilities and implications.

This article was republished by The Conversation under a Creative Commons license. Read the original article.

Photo credit: Natalie B / Pixabay

Leave a Reply

Your email address will not be published.