These notes are designed to help explore the potential of artificial intelligence (AI) tools for making movies. They introduce a process Stephen Quinn has developed which you are welcome to modify (updated 14 April 2024).

Basic versions of the tools listed are free though you will need fast internet and a relatively powerful laptop/desktop and/or tablet. You will need to pay for the advanced or “pro” versions. More details on prices are shown later.

Do remember that the sites on this page are just tools to help you tell stories. Story-telling remains the most important and fundamental skill for all film-makers.

I have had some enthralling experiences with Hume, an AI tool that purports to have a conversation with you and listen like a friend, or perhaps a therapist. Worth a try.

Latest video tools: OpenAI has announced its own text to video tool, Sora. Media reports say you will be able to make videos of up to a minute compared with the few seconds that other tools generate. A date for its introduction has not been announced. Wired magazine has a useful article.

Morph Studio also announced a creative tool consisting of a storyboard boosted by AI. Watch a short explainer and join the waitlist here.

The tool I most want to try is called LTX Studio by Lightricks. Tim Simmons has a thorough preview here. At the end of his video you will find a chance to register for access. This is a 1-minute overview of Lightricks. Here is a video of their launch on 27 March 2024. The video runs 43 minutes and the good stuff starts at 31:21 when director Ari Folman says “I think this is a dream tool.”

Another cool tool is Haiper.ai but at present you can only produce 2 seconds of video from text to video or image to video. The site says lots of other options are “coming soon”. I like the simple interface.

Artflow.ai lets you create characters and worlds. I’m still learning this one but it has a comprehensive range of options so it needs time to learn.

Albert Einstein said: “Creativity is intelligence having fun”. Perhaps we can modify that phrase to “Creativity is having fun with [artificial] intelligence”.

Still images: DALL.E and NightCafe are the simplest tools for creating single images. A fun process involves putting these still images into editing software and then using the Ken Burns effect to simulate movement. I talk more about that below.

Another option is MidJourney though it can be complicated, requires a paid subscription (no free option to tempt you), and works via Discord.

A brilliant newcomer I discovered in November 2023 is Leonardo. The free version will allow you to create several good images per day. Another nice newcomer is RenderNet. I like the simplicity of the interface.

Use MagnificAI to enhance the images you produce. The test results are impressive. But the cheapest plan is USD 39 or AUD 59 a month for the most basic access. Probably better to invest in the TopazAI suite of tools discussed below.

Google offers a free and easy-to-use image creation tool called ImageFX. It produces 4 images based on your text prompt. One cool feature is the fact that each text prompt has a drop-down menu so you can modify your prompt. ImageFX part of a suite of 3 experimental tools that also allow you to create music from text (MusicFX) and play with words (TextFX).

AI tools can also help create cool animations. One of the best is LeiaPix. Here is a tutorial.

Narration/audio/music: Tools like Murf can be used to read and record scripts where you employ stills as the images. Similar tools include Play and ElevenLabs. Until recently I believed humans provided better-quality narration than AI tools.

But ElevenLabs is marvellous. You need to decide for yourself. In late February ElevenLabs announced a tool that converts a text prompt into a sound effect. But you need to join a waiting list at their site to test it. The software also offers a new way to synch AI speech in video for auto translations, which will intrigue the international dubbing market.

Use Enhance Speech, a free Adobe tool, to improve the quality of audio. Their website says voice recordings “sound as if they were recorded in a professional podcasting studio”. Test for yourself.

Stable Audio 2.0 is a cool new tool for creating music and sound effects from a text prompt.

Video creation: In early 2023 these tools were ordinary. By late 2023 they had improved markedly and some of the improvements in 2024 have been wondrous. They offer the ability to create video from text or image prompts. Of the scores of tools I have tested Runway generation 2, PikaLabs and Moonvalley are the best. I will comment on Sora after I’ve trialled it.

PikaLabs is currently free. You can access it through a Discord account or their website. Pika lets you use text prompts to create 3 seconds of video. The prompts allow you to zoom, pan and rotate the video. A watermark is placed in the bottom-right of the screen.

You can also create video from an image you provide Pika. Most people create those images using Midjourney or Fulljourney. Find more information about Pika here.

MoonValley is currently in beta. Like many of the tools it runs on Discord so you need to join that site. I appreciate MoonValley because — unlike most of the others — you are do not need to know their prompt language. You simply click on a series of boxes and select the options you want. This video shows how it works.

MoonValley is currently free and relatively undiscovered, so it tends to be faster than Pika or Runway. Videos run from 1 to 5 seconds depending on whether you choose short (1 second), medium (3 seconds) or long (5 seconds). Image quality depends on the detail in your prompt.

Video defaults to 16:9. Have a look at “best-of-moon-valley” to discover the prompts people chose for videos. A watermark appears on your video or you can choose a monthly subscription which costs USD 7.

I was not impressed with the early version of Runway. But generation 2 of Runway is marvellous and offers a wide range of settings that are easy to use. It runs on both Discord and its own website. This video (duration 1:32) explains how to use the software. Download the videos you like.

Note that output quality is high which means files are large. A 4-second video is at least 4 to 5MB. Once you login you will find two video tutorials on how to use Gen-2 and a recommended Gen-2 workflow.

All video-creation tools offer free versions which let you play with the software but limit video duration and impose watermarks. If you want to make professional-looking movies you need to upgrade to a plan, or find a way to lose the watermark. Hat tip to Caleb at CuriousRefuge for alerting me to this free tool for removing watermarks.

Annual plans are cheaper than monthly, but I suggest trying each tool via a monthly plan before upgrading to yearly.

Mick Mumpitz has created an 11-minute video which introduces all of the main video-creation tools and compares them with Sora. A good introduction to the best vide-creation options. Sora will be available some time this year but no commentators seem to know when.

Kaiber is another good video option. You need to pay for credits. The basic plan starts at USD 5 a month. The free version lets you create up to 18 seconds of video as a way to understand how their AI works. The Pro (USD 15 a month) and Artist (USD 30 a month) plans permit a maximum of 8 minutes of video a month. You need to buy a plan before you can test the software fully.

Pictory also has potential, and offers a 14-day free trial before you need to pay. In late December Google announced VideoPoet, which lets you convert text, images and video into video. Read more about it here.

Another fascinating newcomer is Invideo. It’s a simple way to create an AI-generated video provided you give it a good idea of what you want. Avoid the option where you provide the script because the AI cannot cope with anything beyond the most basic script. This video shows you how it works.

An article published January 2024 about how the movie industry could use AI tools.

Pixverse is an exciting tool for creating video from text or image released in February 2024. Looks highly promising. Thanks to Caleb at CuriousRefuge for the tip.

I’ve had fun playing with StudioD-ID. With the free version you upload a head-and-shoulders photo of yourself, write a 10-word script, choose one of 120 languages and click generate. Then you download the completed video. The free version allows 20 pieces of video, each of about 4 seconds.

And for a bit of fun, try Typpo. Download the app to your smartphone, speak into the app and it converts your words into a video.

As you can see from this example at left, it does not always follow your voice prompts (it interprets fish as goat). And videos are only available as verticals.

Lipdub is a really cool smartphone app that I discovered in mid November 2023. It records a video of people speaking or singing in the language you nominate in the app. It currently appears only available for iOS devices and offers 27 languages.

Lipdub is currently free but inserts a watermark. I tested it in German, French and Hindi and native speakers in all three countries where the language is spoken said my pronunciation was accurate. The example here is in French.

In mid November 2023 Meta (formerly Facebook) announced it was introducing video from text prompts via its Emu Edit process. Read about it here.

In late January 2024 Google announced its text-to-video Lumiere project which will be available later this year.

Tools to avoid: You will encounter poor-quality tools that promise much but fail to deliver. Here I will list them as I find them. Rizzle appears glossy but you cannot edit the videos it makes.

CGI, or computer-generated imagery: This remains a complex and expensive process. But an AI tool, Wonder Studio, lets you animate characters on video that you film. Take a look at the product demo video on the company web site. Lots of other options are becoming available.

Special effects (SFX): Some of the excellent and relatively inexpensive apps available for an iPhone such as Videoleap, Efekt, FxGuru, 8mm, Action Movie and Musemage let you create cool footage that you download into your editing software.

Music: Add music to your videos with tools like Beatoven and Soundraw. Both are easy to use and intuitive but you must buy a subscription to download any music created.

Prompts: The words you choose to describe an image/video you want to create make a big difference in terms of the result. Here is an introduction to prompts, while this is an advanced version. Learning to use good prompts is an important skill. This video simplifies the process and offers a prompt formula (formula starts at 1:02).

Jeffrey Katzenberg: In this three-minute video the former head of Disney studios talks abut how the ability to write accurate prompts will become a “creative commodity against many aspects of storytelling”. That’s business speak for a key skill you need to learn.

A few more cool tools include Ideogram for generating AI images and PrometheanAI for creating virtual worlds. I have been playing with Krea since December 8 and find it fun and easy to use. As with most of these tools you will be prompted to buy a subscription after playing with the free version. You might find a wait-list for Krea.

One possible trick is to use an AI writing tool such as GravityWrite to create the prompts for images and possibly video. Select the film-making tools in the left-hand menu, then choose “AI image generation” from the list. Paste the prompts into one of the video tools above if GravityWrite does not do the job for you. I bought a yearly subscription in December 2023.

Suggestion: AI still images often do not emerge the way you imagined when you wrote the prompts. It might be best to create trial images first based on a draft script, and afterwards rewrite your script to engage with the images that emerge.

Scripting: It all begins with the idea and the script. The human mind is still the best tool for writing quality screenplays. As of March 2024 AI-generated scripts were mundane at best. Many writers recommend Aristotle’s Poetics as a guide to scriptwriting. This video summarises the essentials of Aristotle’s theories and references lots of movie examples.

AI is pathetic for creating humour. The quality is banal though AI can be helpful for routine tasks or prompts via Alexa or similar voice-activated devices.

Research: One possible approach is to use ChatGPT or Microsoft’s Co-Pilot or Google’s Gemini to stimulate ideas or research how others have developed their ideas. In my experience these tools create ordinary scripts.

A fascinating newcomer is Type. This video explains how the writing tool works. It’s free to try, but after a short trial you need to pay USD 29 a month.

I’ve been experimenting with an AI tool for screenplays called PlotDot. In this 11-minute video Tim Simmons from TheoreticallyMedia does a good job in explaining how it works. Looks promising.

New to AI? This article introduces 21 AI tools and would be useful for people new to AI. The blurbs describing each are badly written, probably by an AI tool 🙂

Assistants: Consider using voice-based tools like Alexa to generate ideas and even (bad) jokes. I often ask my Alexa to tell me a joke about (insert subject here) and groan when I hear the result. Useful if you want to offer jokes to grandchildren aged about six.

Unreal Engine has evolved as an exciting option for creating films. The software, currently free, is designed to work best with a Windows computer and takes a while to learn.

This site appears to be the first attempt at using AI to create a TV newscast. Worth a watch.

Watch this video made by The Economist in February 2024 about AI and the future of film-making.

The BBC updated its guidelines about its plans for using AI in late February 2024.

Distribution: This part is your responsibility. Some suggestions are offered at lesson 12 here.

My workflow: This is a suggested approach. Feel free to modify it. I suggest two basic options.

Option 1: This works best with still images. Write a draft script (using FinalDraft 12 or Word or Pages). Use a tool like Play or Murf to record the voices and/or narration via copy-paste from your script. I usually do this on my laptop, then airdrop audio files to my iPad.

I open a video editor on my iPad and insert audio tracks in the appropriate order (I edit with iMovie or KineMaster or LumaFusion, depending on the complexity of the story).

Then insert stills and video to match the audio track/s. With stills I use the Ken Burns effect to create the illusion of movement. I colourise b&w photos with PhotoRevive. Sometimes you need to re-write your script and then re-record the audio and re-insert into your editing software.

To finish, add titles, sound/visual effects and a music soundtrack before uploading to Vimeo or YouTube or a film distribution site.

Option 2: This works best with AI-generated video. Write a draft script. Use prompts from the script to create video with a tool like PikaLabs or Moonvalley. Download each piece of video. This can take time because you are limited to 3 seconds for each prompt. You will get faster with practice.

Then update or re-write your script. Use an AI tool to record your script and save the audio.

Open your editing software, assemble the audio and video, add titles and sound/visual effects and music. Save the completed movie and upload it.

Music: I recommend using AudioNetwork or EpidemicSound, or commission a great musician/composer such as Nigel Thomas.

Please keep an open mind: On this site I’m aiming to show what’s possible and what is currently available. We are very much in the early stages of AI’s development for making movies. New tools are arriving all the time and it is time-consuming to keep up to date.

My AI-assisted movies: My first movie made with AI tools in mid 2023 involved creating still images with DALL.E and NightCafe (see above for details about these tools). I wrote the script and recorded it with my iPhone. I added the images to the iMovie app on my iPad, using the Ken Burns effect to create the illusion of movement. Then I added the narration and changed the duration of video clips to match the words. Finally I added titles and uploaded to YouTube. The project took me a day because I was learning.

For my second movie I used the same AI still-image tools but also included stock video footage. I wrote the script and used ElevenLabs (details earlier) to create the narration. Then I pulled everything together as in the first movie. This project took me half a day.

For my third movie I wrote the script and then used an AI tool called Play to produce a narration (this time with a woman’s voice). I used a combination of PikaLabs and Runway Gen 2 to create video, using prompts from the script. Then I combined the video and audio in the iMovie app on my iPad, added a title and some sound effects and uploaded to YouTube. I removed watermarks by cropping the video during the editing process. This project also took me half a day.

These movies are pretty ordinary. They are simply intended as proof of concept to test the tools.

Stock footage: Two good free sites for stock footage are Pexels (also known by the clumsy name of Canva Germany GmbH, so I stick with Pexels) and Pixabay.

Post production: You might choose to enhance the quality of your completed movies via AI tools such as TopazAI or HitPaw. I had an unpleasant experience with HitPaw when they refused a refund request so I tend to avoid them.

Topaz is powerful but remember that enhancement takes time unless you have a powerful laptop/desktop machine such as the new MacBook Pro with the M3 Max chip.

Resources: I commend to you the website of Neil Chase, a screenwriter who is embracing AI tools for all aspects of the film-making process.

YouTube: I’ve found five excellent YouTube sites that are really helpful, and worth subscribing to.

Caleb at CuriousRefuge produces helpful videos at his YouTube site and an excellent email newsletter which you can subscribe you at the YouTube site.

Mick Mumpitz brings all of the above skills together in his videos. I recommend you subscribe to his YouTube channel and maybe his Patreon site. This video covers many of the AI skills.

Tim Simmons at TheoreticallyMedia provides a range of excellent videos that explain AI film-making clearly and elegantly. He’s one of the best for keeping abreast of latest developments.

Lenny Blonde is also an excellent resource who explains clearly. His channel is called Digital Magic.

Tao (don’t know his other name) offers the fifth helpful channel. These five people will give you more than enough help, and I’m sure you will find others.

Bio of teacher: Stephen Quinn retired as professor of smartphone film-making at Kristiania University in Norway in March 2023 but continues to teach around the world. Dr Quinn’s current research passion is the use of artificial intelligence (AI) tools in the movie-making process. He runs MOJO Media Insights (MMI) which teaches people how to make broadcast-quality videos with only an iOS device. Since 2010 he has taught these skills in 20 countries. From 1975-95 Stephen Quinn was a journalist in four countries with some of the world’s premier media companies (The Guardian, ABC, BBC, ITN, TVNZ, the Bangkok Post, Newcastle Herald, Australia). Between 1996 and 2011 Dr Quinn was a journalism educator in Australia, the UAE, the US and China. He returned to journalism full time from 2011-13 with the South China Morning Post in Hong Kong as digital development editor. Since 2014 he has been based in Brighton in the UK running MMI, a digital consulting company. He makes films when he’s not teaching or writing. Dr Quinn has published 30 books. In the past two decades he has given more than 200 presentations on the future of journalism in 38 countries. Three of his movies are available on Amazon Prime. Details of his books, movies and screenplays can be found here. He also writes and enjoys wine as a hobby. He is hugely proud of his three children and two grand-children.