Another dot in the blogosphere?

Posts Tagged ‘video

Video source

The title of the video makes it sound like the error lay in the artificial intelligent (AI) agent, ChatGPT. The video report even started with a clip of the CEO of OpenAI, the company behind ChatGPT, appearing before a US Congress committee stating how he wanted to work with the government to prevent AI from running amok.

This was a good example of bad news reporting. The agency had already decided to blame the AI for a mistake.

The actual problem lay with the professor who reportedly failed all his students. He supposedly fed their essays to ChatGPT and it responded that it had written all of them.

That is not how to use ChatGPT nor is that a good prompting strategy. Why? Users have discovered how they can make ChatGPT agree with practically anything after it generates an initial draft.

It is not easy to detect AI-generated writing because it mimics a very structured human writer. It is also pointless to cry foul at this point — the ChatGPT pandora’s box has already opened and it has revealed the weaknesses of trial-by-essay and stagnant pedagogy.

The top YouTube video comments hinted at approaches to change pedagogy, assessment, and mindset. One commenter shared how their professor required students to co-write with ChatGPT and to correct its mistakes. This not only recognises its use, it also mandates it and helps students learn to use it strategically and responsibly.

Another commenter suggested that the professor submit their own writing to ChatGPT as a control before accusing students of cheating. This applies the age old principle of “physician, heal thyself”. 

The professor might also reflect on how detecting plagiarism means detecting intent and such a process is multifaceted. Accusing someone of plagiarism is serious — a student who is academically dishonest can be expelled and faculty who do the same can be fired. You have to be absolutely certain if you are going to accuse an entire class of cheating!

The real news is old news: News agencies do not always report responsibly and some professors do not behave professionally. What might be actual news is stories of people unlearning bad habits and learning good new ones. You know, like reporting responsibly and teaching progressively.

This video has a clickbait title, Should we use GMO trees to slow climate change? But I also appreciated how it required me to combine knowledge from my basic degree in biology and my Ph.D. in systemic thinking.

Video source

Cleo Abram wanted to draw eyeballs and earholes to her video with the “controversy” of creating genetically engineered plants to fight climate change. Those that stay, listen, and reflect might actually walk away with this message: It is not just the quantity but the quality of such plants.

Specifically, the project she highlighted:

  • Does not want to introduce trees that are not already part of the ecosystem
  • Wants to improve the carbon capture of trees so that they are more efficient (require fewer to do the same or better job) 
  • Could promote the slow carbon cycle instead of the fast cycle in order to keep carbon in the earth instead of being released

A learner does not necessarily need a biology background to understand this. They need to be scientifically literate to find out why each of those goals are important. But the same person would need to learn to think systemically to understand how the parts contribute to a whole. 

My observation is not simply that people like policymakers need to have these skills. All of us do if we are to weed claims about tree planting and work towards a better shared future. Call that fruit borne from the tree of truth!

Video source

I was entertained and educated as I watched this video by TwoSetViolin. The duo deconstructed movie trailers and suggest how modern ones seem to follow nine rules.

TwoSetViolin's nine "rules" that modern movie trailers seem to follow.

If you give them the benefit of the doubt, their selections and the rules were spot on. If you apply critical thinking, you might wonder if they were being selective in their edits, i.e., they chose specific trailers and parts of those trailers to fit the rules. 

My mind linked this to work-based rules or principles. There are good practices that might emerge that eventually become shared practices. They might even inform policy or become dogma. 

But the same critical light should shine on such rules. Do they work all of the time or just some? When in each case? Why? These questions form the basis of research that should be rigorous and neutral. 

To think and act otherwise is to force patterns where they do not exist and to ignore exceptions to the rules. This is of little consequence when sharing a YouTube video for entertainment. It can have serious consequences when applied to work.

Some people on popular and social media would like us to believe that “kids these days” do not read. This might be true of some kids and a subset of them might have reading difficulties. Those armchair philosophers and critics do not spend as much time talking about that latter group.

What is also true is that there are other kids who love to read, read differently (e.g,. listen to audio books), or read as a result of other media (e.g., movies). I know that I was moved to borrow a digital copy of All The Light We Cannot See after watching this magnificent trailer of its movie adaptation.

Video source 

There was not a single word or voiceover in the trailer. Debussy’s Clair de Lune dominated as the soundtrack and the skilful editing was enough to make me want to read in advance of the movie adaptation.

It is easy to oversimplify or rest on nostalgia about thumbing a book. It is also lazy thinking to make judgements without considering alternatives or nuance. One need only only search archives of old newspapers that reported how reading novels was perceived as a threat [example]. It is that same sort of thinking that will make the current generation judge the next for TikToking or Instagramming. Why make the same mistake?

Video source

What better way than to mark Star Wars Day, May the 4th, with a video about an imaginary alternative to A New Hope?

Video source

Thanks to the video above, I gained some insight to why some people prefer their photos or videos flipped. These are common in phone-shot and vertical format photos or videos that are typically done with the selfie (front-facing) camera.

The video explained that users prefer the flipped photo or video because it mimics what they see in the mirror. They are not used to seeing themselves as how other people see them.

Some might say: To each their own. I say: It is not just about your perspective; learn to take another. Holding up a mirror to yourself is good as long as you recognise that it sometimes distorts and does not fully represent yourself. It takes others to help you see what you cannot.

MKBHD compared AI-powered search in a recent video.

Video source

Bearing in mind that AI is rapidly evolving and some of what he published is already irrelevant, some takeaways for educators might be:

  • Bing provides references more often than Bard
  • Bard shows options to view three other drafts of its answer
  • Bard is faster for basic facts, but less customisable
  • Bing has three levels of answers (more creative, more balanced, more precise) and could be better with complex questions

MKBHD used a strategy that educators can use with students in order to model and teach critical thinking. Test the AI-powered search with something you already are knowledgeable about and compare it with what you already know to be correct. Ba-da-bing, ba-da-boom!

I like watching the comedic offerings of Wong Fu. A spinoff channel, More Wong Fu, recently released a behind-the-scenes video.

Video source

The video started with the team using ChatGPT to write a script for a skit in the Wong Fu style. Therein lay a few reminders that transfer from video production to schooling and education.

For example, the team discovered that they still needed to improve the suggested script with strategic prompts. This was captured in the video and is a reminder that teachers and students cannot simply pass off ChatGPT-generated text as their own — it is simply not good enough and we have to make it ours.

The script also required people with different talents to initiate and create the video skit. They needed people to act, direct, record video and sound, edit the video, etc. People add value to a basic script and do work that AI currently cannot. This is a reminder that technology can relieve low level routine so that we focus on higher order creation or problem-solving.

The middle portion of the video featured an informal contest among Wong Fu staff to sketch a colleague for an animation project. The non-artists among them could have used an AI-powered imaging tool to do this, but they did not. Their efforts were pale compared to their more talented counterparts. For me, this was a reminder of how technology in general is often a leveller (or even a competitive advantage) to those who need it more.

I am still riding the Taskmaster bandwagon and linking what I enjoy for entertainment to what I observe in education.

Video source

This video was ad for Google Translate cleverly disguised as a task. For some, the near instant transliterations and translations that a phone does looks like magic. They call it “magic” because it is wondrous or useful, but they cannot explain how it works.

Skim human history and we learn that earlier humans did the same thing. Anything they could not immediately sense or understand was explained as supernatural, magical, or godlike.

We should know better than that now. Something is new and magical until it is not.

We do not have to understand how a car and plane work to rely on them for transport. We have to know that there are rules and procedures for using them for good. The same could be said about edtech.

In 2018, Simone Giertz turned her Tesla into a truck and called it Truckla. She beat Tesla to creating a truck that actually worked.

Video source

Now Simone is back with a robot that inserts the charger into the truck’s port. She affectionately called it Chargla.

Chargla is a rough prototype and seems to hump the car when it needs to do its job. This makes the process both informative and entertaining.

Some might take away how important it is to work with different and talented people to get a job done. I appreciate how the company she worked with on Chargla relied on two Raspberry Pis and is making the code open source.

Simonela knows that it is not enough to inspire. When learners and innovators actually do the work and perspire, they need access to open resources, not locked up courses or excuses for why things have to be done a certain way.


Usage policy

%d bloggers like this: