« The Screenplay: Scene 17 | Front Page | The Screenplay: Scene 19 »

The Screenplay: Scene 18

Posted by Dave Bull on January 10, 2023 [Permalink]

Dave: OK, I think I see how we might be able to move this forward. Let's try to 'push' it in the direction we might like to go, but keep 'hands off' and let it actually write the thing ... My next prompt gave some background information, and then set a scene:

John (an apprentice printmaker) and his partner Mayu are having dinner with a retired laywer, a friend of Mayu's father, Shinji (who is also a woodblock printmaker). The lawyer is telling them about the time many years ago when he helped Shinji with some difficult legal problems.

The Bot responded with:

* * *

John and Mayu listen intently as the lawyer recounts the story. He explains that Shinji had been sued by a competitor for copyright infringement. The competitor claimed that Shinji had copied their designs and was using them in his prints without permission. After a lengthy legal battle, Shinji was eventually able to prove that he had not copied the designs, and was able to win the case.

The lawyer goes on to explain that as a result of the case, Shinji was able to gain a greater understanding of copyright law, and was able to protect himself and his work from similar situations in the future. He also made sure to always credit any sources he used for his designs, making sure the original artist received proper recognition.

John and Mayu are both impressed by the lawyer's story, and understand the importance of protecting their own work and giving credit to the original sources of their designs. As they finish their dinner, they thank the lawyer for his insight and wisdom.

* * *

[ This Screenplay thread started here. | Previous episode | Next episode ]




Added by: Qaantar on January 10, 2023

Wow, this topic is horribly ironic, given the current uprising that artists are lobbying against AI-generated art and the general issue of large language models NOT citing sources!

But on to two bigger issues -- we've jumped from the "challenge" being a lack of appreciate of their art, to one of credit/copyright. And the format is no longer a screenplay. The latter is fixable, the former, though -- is still really tough without re-prompting all the previous plot points each time...

Still, all in all, a great exploration of the technology, thanks for continuing to share!

Added by: Dave on January 10, 2023

I'm completely open to suggestions on how to play with this. I tried a bit of a different approach because it was just getting repetitive. As for not being in 'screenplay' format, that's because I have been loose with my prompting - sorry about that!

I've also received an email from a person who is not so pleased that I've been playing with this. I was thinking that everybody would understand that I am not in any way taking this seriously. This Bot is simply regurgitating a 'word salad'; I am under no illusions at all that it has any 'understanding' here. But it is definitely making me think more about how we humans actually put together ideas and sentences. I suspect there is a lot more overlap between the two 'systems' than we might be comfortable with!

Added by: Qaantar on January 10, 2023

Ahhh, I've enjoyed your exploration, so thanks again for sharing! I get to deal with this technology frequently at work, so it's fun for me to see how it might apply in other areas.

For me, you've demonstrated both the benefits and the limitations of the technology. I think that we've also been safer here in "fiction land" -- if you tried to make this play a more facts-based conversation on Japanese woodblock prints (such as mentioning life details about Hokusai or Yoshida and such), we'd see the cracks widen even more.

Added by: Dave on January 10, 2023

Yes, there is a clear paradox in this entire concept of generative AI. "Create something new, please. Oh, by the way, don't say anything that isn't true."

Wait, what?

All I can think at the moment, is that once this sort of Bot moves forward into the Search space, or being an Assistant (ala Siri/Alexa/etc.) we would need an 'option switch':

- free style ... anything goes

- verified data only ...

Added by: Marc Kahn on January 12, 2023

Let's flash back to 1950 and Alan Turing's proposal for "The Imitation Game", which has subsequently been named "The Turing Test". I'm thinking that this technology you've been playing with passes the Turing Test with flying colors. I haven't been following developments in AI, so this surprises me greatly!

By the way, do you think that I really wrote the above paragraph?

Added by: Dave on January 12, 2023

A quick parse/search in my 26 years of stored emails (including many from Marc), gives me enough information to think that the phrase "I'm thinking that ..." is indeed a pretty strong indicator that Marc did indeed create that paragraph ...

Added by: Dave on January 12, 2023

A side note here: I just noticed that when prompting the Bot this time around (as shown above), I actually mis-spelled the word 'laywer'. Seems not to have mattered at all ...


Add Your Input


(you may use simple HTML tags for style)

[Due to vast amounts of robot spam, comments with links will be automatically discarded. Other comments may take a bit of time before appearing, until they are confirmed. Thank you for your patience.]