Author Archives: paul bond

“I’d Like To Help You With Your Problem”

I spotted this the other day and noted all the generated imagery, like AI meets CGI or something. It occurred to me that the title is a bit of a sales pitch.

We have a variety of AI pitches this week, covering a range of apps from writing assistants to ChefGPT. So far it doesn’t look like anyone has made an attempt at a hard-sell blackmail style pitch or a poison-pill deal, even though some are admittedly unfriendly to Aggressive Technologies. The main point of this exercise was to look over some of the tools that are out there to see what we might be able to use in the coming weeks. I was curious about AI video, so I played a bit with the Colossyan Text-to-Speech Video Generator.

I had a text generator output the script and used some basic settings in the Colossyan system. It said the video would take about 11 minutes to generate. I had to leave the office after 37 minutes, when it was still processing, but it finished within two hours when I returned. Could I have produced this with my phone and a couple helpers in 11 minutes? Probably. Except then I would have superior voices and more realistic movements.

This is not a comment on anyone’s sales pitch, but rather a reflection on AI generated “art.”

If we look at any aspect of the video – the setting, the character visualization, the animation, the voices, the editing, the background soundtrack – in my opinion, I think we would conclude that they are exceptionally lame. It’s free, so we would expect it to be closer to Blabberize than Ice Ice Matrix, but I’m not sure that the video offers anything more than the script does. On the other hand, one could take advantage of the bland, static nature of their video templates. I remember newspaper comic strips, like Red Meat, would often have static images of two people talking. Something like that could translate into short videos, with bland production contrasting with insane dialogue to humorous effect.

Failed!

I ran into this error message when trying to run Wav2Lip last week:

I did file an appeal, asking what was suspected of being abusive, without expecting a response. Under their usage limitations they list deepfakes.

Anything I’ve made is too transparently faked to be considered deep, but some may argue otherwise. I wonder if it was Bob that busted me. Can’t trust those painter types. It got blocked in some countries due to copyright, so that’s why I’m suspicious. It’s better to work with public domain material anyway though.

I managed to do this with Frank and Vinny. So it should be possible with Grant and Hepburn. I’d have to down load the video, cut the parts before and after the phone conversation, and then separate the two sides to make training videos. After putting words in their mouths I’d have to piece it back together as a conversation. Seems tedious, but possible.

Life imitating AI imitating Art

A few weeks ago we watched the Black Mirror episode, “Be Right Back“, in which a widow has an AI chatbot of her deceased husband. Yesterday I heard that performance artist Laurie Anderson actually has one of these. I guess he wasn’t kidding when he wrote “I’ll Be Your Mirror.”

The idea of characters react to propaganda was a bit of an afterthought last week, but it seems to have worked out well. Some people are using it to find connections, which can be useful as we start to come up with longer narratives and group projects.

We had some nice propaganda too. I especially liked this poster, for several reasons. The letterpress character of the typeface gives it an antique, handmade feel which contrasts to the circuitry in the image. The subtle texturing also gives it an aged feel. But the empty spot inside the image is my favorite part, as it suggests something about the soullessness of the corporate world, or the emptiness of AI. And it plays off of ideas from another post by borrowing one of Dr. Oblivion’s quotes. Of this one, it was said, “This creepy poster will haunt my nightmares forever now,” which may be the nicest comment I’ve seen.

One thing that surprises me is that no one has brought up the environmental costs associated with AI. It seems like a natural for propaganda, and it connects to the broader climate movement. But if that’s not where people want to go, so be it.

Week 8: Take on the tools

2 Replies

We may get a video posted later.

“AI tool laboratory” by net-art generator

3/11/25 – 3/15/24

There are two main projects this week. One is to investigate AI tools. You may use ones you’ve already worked with, or you may look for new ones. We don’t want any duplication though, so we have a form and spreadsheet where you can see what others have chosen. If you look at the list and see the name and URL of your chosen application, then you should pick something else. And don’t use anything that asks for payment or data you don’t wish to share. The other project is to form groups for the upcoming collaborative audio project. Details about the project will come out Week 9, but we want everyone in groups beforehand.

1. AI Tool Laboratory

There are a lot of AI-based tools for creating and manipulating digital media out there. Rundown AI keeps a healthy list which lets you sort out the free ones, and you can probably find many more through Google. Your mission is to investigate a free tool, try it out, and find out its capabilities, strengths and weaknesses. The resulting list will be something to benefit the whole class, ideally.

Aggressive Technologies is looking to expand its AI arsenal, and has put out an open call for acquisitions. They are offering a finder’s fee and/or stock options to anyone who can make a good case for a target, to be awarded if they choose to make the buy.

Have your character (or you if you so wish) make a sales pitch to Aggressive Technologies about why they should acquire it, or why they should put it out of business if you prefer. The pitch can take any form you like, but it should include some sort of example of what the tool can do. Put it in a blog post and tag it AIpitch. To make it easier for everyone to avoid duplicating each other’s efforts, we have a form where you can enter the AI tool you will be using. That form populates the spreadsheet below, so check it first to make sure your tool isn’t already taken.

2. Looking ahead: Prepare for group audio projects

During weeks 9 & 10 we will be creating audio productions for ds106radio. This will be a group project, involving the course theme and your characters. You get to form your own groups. You will be forming groups this week so you can hit the ground running at the start of week 9. Further details on the project will be posted at the start of week 9.

Some advice on group formation:

Get into groups:

Membership: You will have the chance to self-organize into your groups for this project.
Show Ideas: It may be easiest to form groups around show ideas. There are a lot of great ideas out there, so this should not be a problem. You can see everyone’s ideas at https://ds106.ai/category/radioshowideas/ — at least everyone who tagged their posts correctly! The Aggressive Technologies connections may also be a good impetus for forming groups. You can see these at https://ds106.ai/category/aggressiveconnection/. If you see an idea you like, contact the originator about working together. If you have an idea you like, put a call out on your blog and Discord for collaborators.
Use Discord: If you need to find a group, put the word out on Discord in the ds106radio channel that you’re looking for a group to join.
Let Us Know Your Group: I created a spreadsheet to facilitate group formation. You should have received the link to it in an email. Give your group a name, put down a brief description of your show idea, and list the group members. There is also a section for people who are looking for a group.

Group sizes:

Groups should have 3 or more members. If a group grows to 8 or more people, we may decide to split it in two, unless the group can make the case that all members will be actively involved in the show’s production.

Group deadline:

You should start forming groups immediately. Everybody should be in a group by Thursday, March 14, by midnight. Your names and groups should be on the spreadsheet to make it official. If you have not joined a group by that time, you will be putting your fate in our hands. We will assign you to a group, but it will be entirely your responsibility to make the situation work.

3. Daily Creates: Let’s do three again.

4. Commenting: As much as you can

NAGging feeling

Today’s Daily Create asked us to play with the Net.Art Generator. Maybe I should have read the home page, but instead I clicked the Create button as soon as I found it. Moments earlier I had been thinking about AI-generated images and art and Van Gogh’s Starry Night, so that’s what I put in. The output was somewhat interesting to me, reminiscent of Jasper Johns and some of my own adventures in rectangularity. But not good enough. So I tried Starry Knight, just to see what might happen. The text stencilled on the image also reminds me of Johns, but I was able to find the source. To take it one step beyond, and in keeping with the AI theme, I tried Starry Knightrider. I couldn’t decide which to use, so why not all of them?
Now I’m wondering what I might be able to do with KITT and HAL-9000. Apparently MadamBlackWolf thought of this already.

Bot you see is bot you get

When we were kids playing hockey in the driveway, we used to blast this song out of the window and chop our sticks on the pavement in time with the beat. Of course, we were too young back then to truly appreciate such deep and meaningful lyrics. So it’s great that we have sites like SongTell around to give us AI-powered insightful analysis. I just discovered this thing this morning when looking for info on a different song. The thought that something “could symbolize unfulfilled desires or unmet expectations” really resonates here. Is my sarcasm showing?

I asked Dr. Oblivion if the lyrical analysis wasn’t exceptionally lame.

His reaction makes me suspicious. I wonder if I’m the only one.

A frank warning

I wondered about the ethics of working with Bob, so I looked up a list of public domain movies and saw that Suddenly, the old Frank Sinatra assassination classic, was on the list. I downloaded it and skimmed through looking for Sinatra monologues, or anyone’s for that matter. There were a few clips that might be useable, with a little editing. I picked one, then needed something for him to say. I asked Dr. Oblivion how he might warn people about the conglomerate Aggressive Technologies and downloaded his response. I ran it through the Converter App to get a text file, with I fed to ElevenLabs after giving them a sample of Frank’s voice. I ran into some hiccups with the lip syncing script this time. It got hung up on the MP3 upload section and failed 2 or 3 times because something couldn’t connect, but I kept trying and eventually it worked. Then when the video was done, I discovered that the audio was longer than the video, which made the video loop back to the beginning and ruined the effect. I went back to the MP3 file and used Audacity to close up the many pauses. I like the walking towards the camera and talking effect and think it would work well with the right lines, so this works as a proof-of-concept. The key is to have less than 25 seconds of audio so it doesn’t loop. Unless you want it to.

I find it interesting that what he says sounds like something Dr. Oblivion would say, but it doesn’t have the same effect in Frank’s voice.

If anyone wants to puppeteer Frank, you can download the frank.mp4 and use it instead of the train.mp4 with the Wav2lip script.

I’m kinda glad he doesn’t look at the camera.

I was glad to see someone successfully took on the Talk to the Bot assignment. A few others tried but couldn’t get it to work. As I watched the voice-over with avatar part of the video, I wondered what would happen if you tried to lip-sync a still image. At the tail end of a previous exercise, I noticed that the program animated the Aggressive Technologies logo, so I thought it might work. I uploaded a Van Gogh self portrait to iMovie and stretched it out to 10.6 seconds and shared the file as an MP4. That gave me part of the puzzle. I went to ChatGPT and input “if you could see what I hear” and copied out its response, then pasted that in ElevenLabs to generate a voice. I don’t know what Van Gogh sounded like, so I used a different painter. I took that MP3 and used the Wav2Lip script to combine it with the image. It worked, but it’s a little weird with moving lips on a still image.

I recalled that there was an online tool that would animate still photos in weirdly creepy ways, so I gave that a try. Sometimes it refuses to animate things that aren’t photos, but it worked this time. I downloaded their video and ran it through the lip-syncer. I can’t say that it’s less weird, but at least he blinks. That’s a little more human, right?

I’m kinda glad he doesn’t look at the camera. Based on this experimentation, I may need to revisit my Right Said Freddy Kreuger Mercury video.

Week Seven: “Ask not what AI can do for you…”

Propaganda. N. From the Latin propagare, meaning “to go viral.” A communication intended to influence or persuade, to further an agenda, usually through an emotional appeal rather than a rational one.

What is your character’s position on AI? Aggressive Technologies? How does it impact them and the people and things they care about? Would they promote or protest?

There are many aspects we could think about. Consider employment, environment, advancement, economy, ethics, surveillance, personal data, personal image rights, voice rights, control and ownership of data and information, copyright, how to protect people from abuses, how to respect people’s freedoms and intellectual curiosity and creativity, consolidation of power … what else could you add? We’ve looked at movies and manifestos that touch on some of these ideas. Now is the time to take a stand.

We will use this week to create propaganda. The term is generally used in a negative sense, but emotional or persuasive communications for positive ends can also be considered propaganda. This means, as usual, that you have a lot of leeway in what you create. It could be propaganda that your character supports and shares, or it could be propaganda that your character encounters in their daily life. It should relate either to artificial intelligence or Aggressive Technologies.

Design some propaganda. This could be a poster, postcard, bumper sticker, web meme, or something else. Be sure to use design elements like imagery, type, color, shape, etc. to convey an emotional appeal

Examples: Keep Calm and Carry On

Examples: Rosie the Riveter

Do not feel limited by these examples. They’re just the first two things I grabbed from Wikipedia. Put your design in a blog post with the usual write-up, but also tell the story of how it connects with your character. Tag the post propagandadesign.

Create a Public Service Announcement: This could be audio or video, but either way should both voice and background sounds. The announcement should spread a message about AI or Aggressive Technologies, or both.

Examples:

Again, do not feel limited by these examples. They’re just the first things I found in the Internet Archive.

Good ways to find background music that you can use without running into copyright issues are to use CC Search or to google open source music. Embed your PSA in a blog post with the usual write-up, but also tell the story of how it connects with your character. Tag the post PSA.

Maybe they express what your character feels. Maybe they show what your character encounters

Complete 3 Daily Creates: You must complete at least three daily creates this week. Make sure you also blog your TDCs.
Commenting: In addition to the usual commenting, do a blog post about how your character would react to a few of the designs and PSAs that the rest of the class produces. Make sure your post links to the source posts so people get pingbacks. Tag this post reaction.
Weekly Summary: You all know what this is.

A word on uploading videos

It looks like a few people have uploaded videos to their Wrodpress media libraries. It is probably better to upload to Youtube or Vimeo instead, and embed videos in your posts from there.

If you go into cPanel and look at your usage statistics, you can see how much drive space you have used up and how much is left. Videos take up A LOT of space. The last 3 minute video I made was about half a gigabyte. By uploading it to Youtube, that half a gigabyte sits on google’s servers, and the only thing taking up space on my site is the URL to the video, which is only slightly larger than zero. People may have concerns about having their videos in a public space like Youtube. You can mark videos as “Unlisted” during the upload process, which means they will not show up in a search or be visible to anyone who might find your channel. They would only be findable through the URL, or on your blog. You can also delete them or make them private later.

Tech Noir

The Darker Side of ds106