Day 4 Creating videos
Welcome to Day 4 of the 12 Days of AI. Today’s focus will be looking at how AI can create videos and how they could be useful in your own practice.
What is Hour One?
Hour One is a AI platform that enables you to convert written text into videos. It has lifelike virtual characters that are modelled on real people to read out the written information you have inputted into it. Hour One’s privacy policy.
How do I access it?
Go to Hour One and select ‘Create your free video’. It will ask you to create a new account with HourOne. We have listed alternatives below you may wish to use instead.
Your task:
Create a video on any Christmas, Chanukah or winter holiday/seasonal related topics. Here is one we created explaining How #12DoAI Works that you may have already seen and given feedback on.
Here are a few ideas:
- The Icelandic Yul cat
- Favorite Christmas Films
- Best Chanukah dishes
- Top 10 winter-themed songs
- Seasonal bake off!
Alternatives to Hour One:
Optional reading:
- The Dangers of AI In Video Production (Medium – free account required)
Join the conversation:
Reply in the comments below or post on X using #12DoAI. Be sure to post a link to the video you created.
- How did you get on with the task – did you find it useful?
- How easy was it to create your Hour One video?
- Can you see any advantages of using Hour One with their virtual characters to create your videos – rather than recording yourself doing them? Disadvantages?
- How could you use these videos in a teaching context?
- How else could you use these AI generated videos?
- Can you see dangers of creating AI generated videos?
Join the competition
We’re running a competition during this course where you can win actual prizes! Learn more aboutt how to enter the competition.
Wow! Well that was a marathon task!
Maybe it’s my slow hotel wifi, but that took forever to render.
HourOne is easy enough to use, though it stll took me a few takes to get all the text and image edited as I kept missing bits – definitely a need for close attention to detail to get it right first time.
It also highlights the need for pre-planning and designing content – You can’t just wing it and go with what you’ve got or you finish up with rubbish text and random images 🙂
Greater variety in accents would be nice, and if I was going to use this a lot I would likley look at a paid subscription to be able to clone my voice. Though timewise, I could have done a selfie recording much quicker.
Here it is:
https://player.hourone.ai/98e7821784d24f54839a5df3d06c37fa
Well done on completing the task Sue! Yes it is a longer task and you could spend a lot of time on creating videos – hopefully this have given you a taster of what can done! Maybe we need a Australian accent too!
absolutely agree Sue. I found it challenging but that could be because it was the first time I have used a platform like this!
Nice one, Sue! I haven’t had a go at this task but this has peaked my interest. And yes, we need Aussie accents! 😉
Merry Christmas!
I used Synthsia and was very easy to use but as its a free account had limited options to edit/change in terms of speaker etc and limited word count. Its impressive in that it has created a video but it feels very impersonal and I can imagine will become very generic feeling. Its a bit obvious to say but the human touch is missing. I could see how this may be useful in certain industries but at moment not in education/teaching. Maybe as an instruction on how to use something or solve a particular issue. They will of course improve and combined with voice impersonation and facial impersonation it could have quite wide ranging problems socially.
HI Andrew I’ve not used Synthsia before – does the free version allow you to share the finished video?
does this link work? https://share.synthesia.io/0951de7b-8722-4740-85f4-7dcb1bcf945a
it offered a share button to twitter, facebook, linkedin and an embed code
Yes it works thanks Andrew….the face is more ‘expressive’ compared to the Hour One i did!
I used one hour and as the name suggest it took an hour to understand the platform.
I didn’t like the American accent I suppose there is a way to change that?
it was too generic, impersonal and it was very Max Headroom for anyone who remembers that show? it also looks ‘cheap’? did anyone else find this ?
the hands on my ‘elf’ were very robotic. As Sue mentioned I could have shot a video a
lot quicker on my iPhone.
will I use again doubt it.
apologies I typed the wrong name I for this AI app ‘hour one’.
You can select other voices. I picked one with an English accent.
Creating a script and scenes seemed a lot of hassle – many of the videos I create myself are ‘off-the-cuff’ and ‘straight to camera’ . Discussing something like Christmas films made the lack of emotion in the voice very apparent and slightly disconcerting. But the main thing about many of the videos we created is that they are part of a blended model – so we want to presenter to be ourselves, so as to have maximum credibility with the leaners – It would be more us for us to use a text based AI to refine and polish the script I guess.
I thought this was pretty good after working on changing the character, trying to match the voice, bit of background music and entering my own PowerPoint (I made a 2 slide presentation not Christmas themed…sorry not there yet 🙂 )
It would probably be better suited to a commercial presentation selling something and not really what I do but I was impressed enough. The presenter was very stilted but I chose a non-smiling one. The rendering, as mentioned, takes ages so putting something together, at the moment, takes a lot of time. I thought there was a reasonably diverse choice of presenters and voices although some were quite ‘demanding’ sounding voices that I wouldn’t want to listen to for longer than a minute. The layout was similar to Adobe Spark, if anyone has made videos with that, and it’s something you could waste time on perfecting for hours. As I’m a small team I would just use myself as a presenter if using in videos for students as I would want to personalise but in other industries I feel this would be useful.
Created a video on ‘Forming a position on AI safety’ – it was was a surprisingly useful process in being a different way of writing. This was an unexpected positive – another way of getting beyond a blank page! Of course, this is the beginnings of this tech and there’s lots of rough edges in the video below! Maybe it would be useful to explain this isn’t fully developed tech and will improve? Anyway, an attempt to address one of the set tasks in a 2 minute video:
https://player.hourone.ai/9f0e490ef2b84c1d9700a3b8246463ff
I would like to echo everything Sue said, it was only after my second attempt did I realise you could upload your own background images, but I still made a mistake. I think the name is fitting LOL because it literally took me around that to complete the task. I also think the accents could be better. I think with a bit of practice I could use it for introduction to online teaching, however it would be useful for those watching to have either subtitles or transcript, maybe that’s available with the paid subscription. I will add that I got Bing co-pilot to write the script.
Here’s my video: https://player.hourone.ai/0326ab2c43754087a1477346aafcbaf0
I found it fairly easy to create an Hour One video, but difficult to make it good!
I’m not a fan of being on camera, so using this could be a reasonable alternative. However, the characters I tried are a bit uncomfortable to watch. They seem to bounce their head awkwardly when talking and they stare at you constantly without blinking. Creepy!
I noticed that the voice transition between slides was too abrupt and tried the pause feature. It didn’t seem to make any difference in the preview, but once rendered, it was much better.
The AI-created speech was quite generic, and I suspect some elements weren’t entirely accurate. I think I would prefer to use my own words if creating another video, rather than risk sharing misinformation.
It also takes a while to render the final video, so you wouldn’t be able to post anything immediately online.
I think these would probably be fine for short explainer videos, demos etc, but for something longer and requiring a bit more personality, I would look elsewhere.
It’s quite a methodical, by-the-numbers tool which I found useful but as others have said it takes some time to make this a polished product, which can test your patience. The options for different parameters are quite good but the voices are a bit unnatural and not particularly engaging. I also didn’t really like the way you have to refine your content by answering some marketing questions at the beginning, which limited the type of video I could create (unless I was doing something wrong). I can see this as a useful tool to create quick content for students but I would need to do a lot more research and practice to get the most out of this tool.
A really interesting tool that I didn’t know about before. As others have said it would take a while to write a script and tweak the options to get a good result, but I can see its potential.
I found HourOne quite straightforward to use, and it had a guide to demo some of the extra buttons. It did take a bit longer than a text-to-speech tool to render video, but not too long, considering.
It has a number of ‘English’ accents, but was disappointed not to see a Welsh, Irish, Scottish or other northern accent, and was a bit taken aback with the ‘Clone your voice’ tool (paid), wondering if this could pose security risk somehow.
I agree with Coral, that such AI-created videos could be useful for short explainers.
However, for an intro to an online course, a more natural recording of the tutor or even an audio with images and tutor’s voice would be more engaging, I think.
English language students may find AI video creation useful for telling stories or practising presentations, as I thought the intonation was better than some text-to-speech tools.
My very short attempt: https://player.hourone.ai/17e0d07c5d1a415ea4033cf72265dd6c
There seems to be an Elves theme going on here Helen 🙂
I found HourOne quite straightforward to use, and it had a guide to demo some of the extra buttons. It did take a bit longer than text-to-speech tools to render a video, but not too long, considering.
Was disappointed that there are a number of ‘English’ accents, but none sounded northern or Scottish. The Irish voice was free, but Welsh not free.
I was a bit taken aback with the ‘Clone your voice’ tool (paid), wondering if this could pose security risk somehow.
I agree with Coral, that such AI-created videos could be useful for short explainers.
For an intro to an online course, a more natural recording of the tutor or even an audio with images and tutor’s voice would be more engaging, I think.
English, and other language, students may find AI video creation useful for telling stories or practising presentations, as I thought the intonation was better than some text-to-speech tools.
My very short attempt: https://player.hourone.ai/87406d3f24304a39b19d4aa7ea895ab4
Great challenge. I asked the generator to create a script on winter poetry, so the blurb about winter and writing were AI generated. I added in the poetry to see what AI voices could do with it. Not sure if, with the right level of understanding, I could control the intonation a bit more. I wanted to embed a hyperlink to a contrasting video, but I was unable to work out how to do that, but maybe with the full version, this would be possible. I see that you can embed a screen capture, so perhaps that would be the answer. Didn’t quite have enough time to do that. I’m a bit behind with the challenges for the 12 days, so I was reassured by your AI presenter in the demo video 🙂
I asked OneHour to come up with 3 scenes about winter traditions across the world – no idea if the info was correct unless I check it (do they eat buckets of KFC at Christmas in Japan?!!). The image selection could have been better, but I was pleasantly surprised by the Scottish accent which didn’t sound too robotic, as some of them do.
I think the use for teaching would depend on the context of what the video is about. I think that students hearing our voices is one of the ways we build a relationship with them. As for seeing a person talking on the screen, I thought that research had shown there was no benefit to this (Mayer’s principles ) but I’m happy to be corrected.
I agree that HourOne is easy enough to use. Figuring out its optimal use might take some time, but it offers an entertaining initial experience of creating AI-generated videos! It is difficult to assess its use in academic environment as it is rather impersonal and robotic. Potentially, over time, it could generate highly convincing fake news!
Here you go with a very quick try!
https://player.hourone.ai/ef79c0e9581341cd9df110f62ea4ffc0
I will need to come back to this one. I find it hard to create new accounts… Would love to be able to test some of the tools without having to create an account before committing. What do others think?
Yea I know what you mean Chrissi all these tools require you to create a new account. I have a separate email to sign up for all these new sites.
The separate email is a great idea. I did think about it and then was too lazy to set it up.
I am with you on this. Another account to keep a note of etc.
Wow, it was relatively easy and I can see how to improve it. The script needs to be sharper and I should have had a final summary scene with the presenter. For a first draft, it’s ok!
I drafted some text about Advent wreaths and used the basis style/palette. I found Advent wreath images via the image tool, using stock images from Unsplash. I didn’t check if they were copyright free! I had taken a photo of my Advent wreath but it didn’t upload, it didn’t meet the image requirements. I used an older presenter to represent me and an Irish voice, Emily. There was the option to clone my voice but I didn’t take it. I would need to know who would have access to it. I set up four scenes with various layout and then created the video. The video took about 10 minutes to create / render. It used 1 of three free minutes.
The video is at https://player.hourone.ai/d28483d2c172420686cc07bff7f51bf4.
I am in two minds about this. I can see how it might have value in a teaching context for explanations etc but is it better than simple to camera videos? I have more questions than answers but I enjoyed the task and would explore this more. Might also be worth exploring the other options.
I used Claude to create a script about sinister Christmas Traditions and then used Hour One to generate it. I think the scenes with just the narrator voice but without the image are a bit more natural but they are very static.
https://player.hourone.ai/fa7c92096aa34851bc62ae261aec50a8
I tried Hour One the first time and I think the AI generated voice over is much better than those I have been using in Articulate Storyline. If you know how to create video, you would know to break down the script into paragraphs and add different scenes to the video rather than just the presenter. It is very efficient to create an intro video with the presenter and free stock media provided in Hour One but I found it quicker to just download from Pexel or other free stock media site and upload to Hour One, add it to the scene and change the layout and colour of the scene.
I am impressed with the speed and quality to create an intro video for free!
HourOne reminds me of simpleshow, an animated video making tool I came across several years ago. It is great to make bite-size videos (with a free plan). Such videos would be especially suitable for a quick instruction, introduction or explanation of something. While it took quite a while to get the video ready, it’s intuitive to use and its interface is familiar. I made a short introduction to Academic Support, and the result to me was amazing – very clear. However, it looked a bit too institutional, lacking some quirk and human touch. All in all, I find the tool very useful.
Someone I know was writing research paper on the use of these technologies in teaching, and the feedback was generally that full photorealistic generated personas were received much less favourably than simple generated line drawing animations. Those obviously use some similar AI, but step away from uncanny valley territory. That’s a bit like what I mentioned the other day, preferring drawings of generated humans from the image tools like Dall E, rather than photo-realistic outputs.
I found creating a quick video quite simple and got pretty good results, so suspect it might depend on which avatar and voice combination is chosen (my character blinked normally and lip-synced well with the words). From viewing a few of the other examples, some other avatars do seem a little wooden and creepy.
Although they’re not 100% realistic (yet…) I do think this kind of video creation has an application, and although viewers will know it’s not a real-person, I don’t think that’s a bad thing necessarily.
Personally, as a stammerer, I hate talking to camera, so would definitely utilise this kind of tool for creating talking-head type videos with my words. For explainer videos whose content may need to be changed quickly and easily without needed to edit or create a brand new video, this tool seems great.
as a budding DJ in my spare time, I thought I’d have a go at creating a spoof training video – not quite adhering to the brief, although to be fair the presenter is wearing a Santa hat!
https://player.hourone.ai/937e097f164d4eff885925f161b20324
As others have commented, it is very time-consuming to produce a video, this is partly down to rendering times ( which I understand it’s much faster if you upgrade to a paid plan).
in terms of advantages of using virtual characters, clearly this is much faster, less labour intensive and cheaper than recording yourself. As it stands currently, it’s a bit of fun, quite gimmicky, maybe good for producing corporate videos, but without the authentic personality inherent with a real person I think it has limited use in an educational context. Having said that, the technology is progressing rapidly and I imagine you’ll soon be able to upload a photo of yourself and AI will do the rest. The paid for version can already take a recording of your voice and supposedly mimic that element, so perhaps creating a video with just a VoiceOver is an option for now?
I want my money back Darren!
🙂
I created a video by providing Hour One a very basic scriptof my job title, pronouns and a short description of what the tasks involved in my role. This was a very easy process, as I came up with the script and within some time Hour One had a 1-2 min video ready, with a slideshow, a person and voice over. Very impressive.
What I’m unsure about however, is that Hour One fleshed out my basic script and added words in the video, which were not there. I understand that this is how AI generates the content and it was useful, but I didn’t like the language used. I found it way too formal, as it advertised my services almost like a sales pitch ‘Antonis will help you succeed in your academic endeavors’ etc., which I felt went against our friendly approach at UAL libraries.
The autogenerated text is a bit dangerous, as Hour One might end up misinforming the students about what I can help them with, and it might also give them unrealistic expectations as to what they can expect from me as a Librarian. For that reason, I’m not sure that I’d trust using Hour One for my teacing just yet. Also, I couldn’t afford to even if I wanted to.
Day 4 of the 12 Days of AI: VIDEO Summary !
https://totallyrewired.wordpress.com/2023/12/11/day-4-of-the-12-days-of-ai-summary/
I thought this task was pretty straightforward however I thought it took a long while to process and render the final output. I can definitely see the benefits of how the video can be used in a teaching context. from my perspective I thought this could be quite interesting to use in promotional material of advertising a live brief or providing an exciting summary of an upcoming projects that I deliver here at UAL. I wouldn’t be keen on doing something like this immediately however it might be worthwhile coming back to this when they have worked on it a bit more and developed the software.
I do think creating videos such as these two take away the human approach which therefore makes it harder for participants/viewers to relate to what you’re trying to do.