The Only Winning Move [Short Film: In Nature]

When I started writing and planning my stop-motion short In Nature, I made a decision not to use AI anywhere in the production, for several reasons:

• I want to enjoy the physical processes of model-building and stop-motion.
• While I want to make the best film I can, I also enjoy the process of trying things, making mistakes and learning from them.
• The flaws that do still appear in the film are the scars from this trial and error process. They mean something, and it's OK for them to be there.
• I have friends in different artistic or creative fields and I know that the increasing use of generative AI is a real anxiety and a genuine threat to livelihood for many people.
• I am aware of the use of AI-generated Lego-style animations for the purposes of propaganda in certain global conflicts and don't particularly want my short horror film to be accidentally mistaken for propaganda.

I think my main point is that there is a creative journey that can be just as important, and just as enjoyable, as the destination, and using AI robs me of that journey.

I'm not criticising people for using AI in films. It's inevitable that it will be, and I might do so in future. While I think I have succeeded in my goal for In Nature, I'm aware that it's becoming harder to avoid it, not least as AI tools are being gradually integrated into art, design and editing software - over time this may well take the place of processes such as bluescreen. It would be a shame if past skills and techniques were completely lost. In a slightly different context I remember being impressed by Disney's decision to make The Princess And The Frog using traditional animation techniques rather than relying on the CGI that has been used (admittedly very well) in their other films.

I'm also not saying I never use AI. I'm definitely ai-curious and enjoy playing with AI and discovering what it can and can't do - I've blogged about these experiments before and there's another one to come shortly.

Living through the Singularity is turning out to be an interesting experience - even more interesting than living through the dawn of home computing or the Internet. And yes, I am that old. I'm still fascinated to find that we have created flawed AI that can hallucinate, miscount, behave emotionally, irrationally or criminally, rather than something with superhuman intelligence, 100% accuracy and a sense of ethics - we are, quite tragically, building Samaritan rather than The Machine.

Zoom And Enhance

For many years now, fans of TV crime drama and sometimes science fiction have been taking the piss out of episodes featuring image enhancement. You know the trope: "stop, that frame, now zoom in, can you enhance that?" and four pixels are sharpened into a UHD quality portrait. Supposedly there are algorithms that can do this, and by coincidence this is something the geekiest member of the CSI:Your Town Here crew has been working on in their spare time. A selection of particularly amusing examples has been compiled by Duncan Robson in the YouTube video above.

However the game has changed. Until recently we could simply laugh and say that no such algorithm exists. Now we know what the algorithm is - it's Stable Diffusion, the principle used by Midjourney, Dall-E and other generative art AIs. And it's no longer impossible - instead it's dangerous. AIs are prone to bursts of creativity and "hallucinations" (the term used by AI scientists to mean "lying"). When an AI enhances an image, the face that is digitally pencilled in doesn't have to be the actual face of the suspect seen on CCTV, it just has to be a plausible image. It could be anyone's face. It could be your face.

So here's my question: how many innocent fictional people in episodes of CSI have been wrongly convicted of the most heinous crimes on the basis of an AI-enhanced image?

JUSTICE FOR FICTIONAL IMAGE ENHANCEMENT VICTIMS NOW

Firefly Season 2 Episode Guide

 CONTAINS AI-GENERATED CONTENT

**Episode 1: "The Heist"**

Zoe receives a job offer from an old war buddy to steal a valuable artifact from a heavily fortified museum. However, when their plans are foiled by unexpected security measures, Zoe must rely on her training and quick thinking to save the crew and complete the job.

**Episode 2: "The Lost Colony"**

Jayne receives a message from an old acquaintance on a remote planet that was once a thriving colony. When they arrive, they find that the planet has been abandoned for years, and they must unravel the mystery of what happened to the colonists.

**Episode 3: "The Huntress"**

Inara's past comes back to haunt her when she is hunted by a former client who has become obsessed with her. In order to protect herself and the crew, Inara must use all of her skills as a Companion to outwit the deranged client.

**Episode 4: "The Diplomat"**

Simon is recruited by a group of diplomats to mediate a peace treaty between two warring factions. However, when the negotiations break down and violence erupts, Simon must put his medical skills to the test to save lives on both sides of the conflict.

**Episode 5: "The Genius"**

River's unique abilities are put to the test when the crew is hired to protect a brilliant scientist who has developed a groundbreaking technology. However, when the scientist is kidnapped, River must use her intellect and intuition to track down the culprits and rescue him.

**Episode 6: "The Outlaw"**

When Mal is captured by a group of bandits, he must use his cunning and charm to talk his way out of trouble. However, when he discovers that the bandits have a personal vendetta against him, Mal must rely on his friends to mount a daring rescue mission.

**Episode 7: "The Ghost Ship"**

Serenity encounters a derelict spaceship adrift in space, and they investigate to see if there are any survivors. However, they soon discover that the ship is haunted by the ghosts of its former crew, and they must confront their own fears and regrets to escape the ship alive.

**Episode 8: "The Traitor"**

When Serenity is hired to transport a group of Alliance soldiers, they discover that one of them is a traitor working for a dangerous criminal organization. As they try to uncover the traitor's identity, they find themselves embroiled in a deadly game of cat and mouse.

**Episode 9: "The Reunion"**

Kaylee is reunited with her estranged father when Serenity is hired to transport him to a distant planet. However, their reunion is complicated by the fact that Kaylee's father is a wanted criminal, and they must decide whether to turn him in or risk their own freedom to help him.

**Episode 10: "The Final Battle"**

In the season finale, Serenity must defend a peaceful planet from an invasion by the Alliance. As they face overwhelming odds, each member of the crew must confront their own fears and doubts to fight for what they believe in.

This episode guide was generated through discussion with ChatGPT. I had to establish that I was asking for a hypothetical episode guide for Series 2 had it been made, and then in response to the first few
attempts I encouraged ChatGPT to avoid repetition and for some episodes to revolve around individual characters. Everything else is ChatGPT. 

Real life sci-fi: advances in computing

On a snowy day a couple of years ago I was waiting at a station for a train no doubt delayed by adverse weather. The station's computerized announcement system relayed changes in the estimated time of arrival at regular intervals. Finally, in the same spliced-together female voice it made the following announcement: "On behalf of South East Trains I would like to apologise for the cancellation of the 10.20 train to Tonbridge" (The train details have been changed).

I can think of a number of explanations for this curiously personal announcement:

1. The computer is indeed responsible for scheduling trains and is taking responsibility for the decision, made by itself and possibly overruling human advisors, to cancel the train.

2. The computer is acting as a kind of confessor, having taken responsibility for the decision from its' human colleagues at South East after they have confessed their sins to it.

3. The computer, while not actually having taken the decision, has gained awareness of the inconvenience for its cold and frustrated passengers, and has become guilt-struck.

Any of the above imply a computer with an impressive level of emotional intelligence and self awareness way beyond anything I've heard of outside science fiction. Conversely, it hasn't escaped my attention that the computer might simply have been programmed to speak in this grammatically and contextually inappropriate way. In this case, if the apology is automatically generated or someone has pushed the "Apologise" button on their keyboard, but no apology has actually been spoken, does it count?

This event reminded me to stay on the lookout for signs of computers insidiously taking over the world. Today I came up against another computer - I was sent an inconvenient delivery time for a new desktop, and when I phoned a call centre in Limerick to rearrange it I was informed, by a woman with a beautiful but fear-tinged voice, that the decision to deliver at this time had been taken by a computer, and that there was no-one in the entire delivery company who had the authority to overrule it. With the UK news agenda dominated by Wossygate and Sargentgate I may have missed coverage of a robot revolution in the Republic of Eire.