Tag Archives: Artificial Intelligence

AI, Content and Context

Leave a reply

I found this quote very interesting. On his podcast, Diary of a CEO, Steven Bartlett is talking to Daniel Priestley and Steven mentions that Open AI’s Sam Altman believes we are not far away from a 1 person company making a billion dollars, using AI rather than other employees. Daniel pushes back and says while that might happen, a more likely and more repeatable scenario would be a 5 person team. Then he says this:

“AI is very good at content but not context. And having 5 people who share a context and create a context, together… then the content can happen using AI. AI without that context, it doesn’t know what to do, so it doesn’t have any purpose.”

Daniel Priestley

Like I shared before, “The true power and potential of AI isn’t what AI can do on its own, it’s what humans and AI can do together.”

This idea of context versus content seems to be the ingredients that make this marriage so ideal. This is noticeable when generating AI images, as I’ve done for quite some time, creating images to go with this blog. For example, I’ll describe something like a guy on a treadmill and maybe one of the four images created would have the guy backwards on the treadmill – content correct, but not context. As well, AI is really unaware of its’ own biases that humans can more easily see. These context errors are common.

But just as AI will be better teaming with humans, humans are also better when they team with other humans, rather than being solo. We miss context too, we struggle to see our own biases, unless we have people around us to both share and create the context.

The best innovations of the future are going to come from small teams of people providing rich contexts for AI. And while AI will get better at both context and content, it’s going to be a while before AI can do both of these really well. It’s what AI and humans can do together that will be really exciting to see.

Content Free Learning (in a world of AI)

Leave a reply

Yesterday, when I took a look at how it’s easier to make school work Google proof than it is to make school work AI proof, I said:

How do we bolster creativity and productivity with AND without the use of Artificial Intelligence?

This got me thinking about using AI effectively, and that led me to thinking about ‘content free’ learning. Before I go further, I’d like to define that term. By ‘content free’ I do NOT mean that there is no content. Rather, what I mean is learning regardless of content. That is to say, it doesn’t matter if it’s Math, English, Social Studies, Science, or any other subject, the learning is the same (or at least similar). So keeping with the Artificial Intelligence theme, here are some questions we can ask to promote creativity and productivity in any AI infused classroom or lesson:

“What questions should we ask ourselves before we ask AI?”
“What’s a better question to ask the AI?”
“How would you improve on this response?”
“What would your prompt be to create an image for this story?”
“How could we get to a more desired response faster?”
“What biases do you notice?”
“Who is our audience, and how do we let the AI know this?”
“How do we make these results more engaging for the audience?”
“If you had to argue against this AI, what are 3 points you or your partner would start with?”

In a Math class, solving a word problem, you could ask AI, “What are the ‘knowns and unknowns’ in the question?”

In a Social Studies class, looking at a historical event, you could ask AI, “What else was happening in the world during this event?” Or you could have it create narratives from different perspectives, before having a debate from the different perspectives.

In each of these cases, there can be discussion about the AI responses which are what students are developing and thinking about… and learning about. The subject matter can be vastly different but the students are asked to think metacognitively about the questions and tasks you give AI, or to do the same with the results an AI produces.

A great example of this is the Foundations of Inquiry courses we offer at Inquiry Hub. Student do projects on any topics of interest, and they are assessed on their learning regardless of the content.  See the chart of Curricular Competencies and Content in the course description. As described in the Goals and Rationale:

At its heart inquiry is a process of metacognition. The purpose of this course is to bring this metacognition to the forefront AS the learning and have students demonstrate their ability to identify the various forms of inquiry – across domains and disciplines and the stages of inquiry as they move through them, experience failure and stuckness at each level. Foundations of Inquiry 10 recognizes that competence in an area of study requires factual knowledge organized around conceptual frameworks to facilitate knowledge retrieval and application. Classroom activities are designed to develop understanding through in-depth study both within and outside the required curriculum.

This delves into the idea of learning and failure, which I’ve spoke a lot about before.In each of the examples above, we are asking students challenging questions. We are asking them to critically think about what we are asking AI; to think about how we can improve on AI responses; or, to think about how to use AI responses as a launching point to new questions and directions. The use of AI isn’t to ‘get to’ the answer but rather to get to a challenging place to stump students and force them to think critically about the questions and responses they get from AI.

And sometimes the activity will be too easy, other times too hard, but even those become learning opportunities… content free learning opportunities.

Google proof vs AI proof

1 Reply

I remember the fear mongering when Google revolutionized search. “Students are just going to Google their answers, they aren’t going to think for themselves.” Then came the EDU-gurus proclaiming, “If students can Google the answers to your assignments, then the assignments are the problem! You need to Google proof what you are asking students to do!”

In reality this was a good thing. It provoked a lot of reworking of assignments, and promoted more critical thinking first from teachers, then from students. It is possible to be creative and ask a question that involves thoughtful and insightful responses that are not easily found on Google, or would have so few useful search responses that it would be easy to know if a student created the work themselves, or if they copied from the internet.

That isn’t the case for Artificial Intelligence. AI is different. I can think of a question that would get no useful search responses on Google that will then be completely answerable using AI. Unless you are watching students do the work with pen and paper in front of you, then you really don’t know if the work is AI assisted. So what next?

Ultimately the answer is two-fold:

How do we bolster creativity and productivity with AND without the use of Artificial Intelligence?

This isn’t a ‘make it Google proof’ kind of question. It’s more challenging than that.

I got to hear John Cohn, recently retired from MIT, speak yesterday. There are two things he said that kind of stuck with me. The first was a loose quote of a Business Review article. ’AI won’t take over people, but people with AI are going to take over people.’

This is insightful. The reality is that the people who are going to be successful and influential in the future are those that understand how to use AI well. So, we would be doing students a disservice to not bring AI into the classroom.

The other thing he said that really struck me was, “If you approach AI with fear, good things won’t happen, and the bad things still will.”

We can’t police its use, but we can guide students to use it appropriately… and effectively. I really like this AI Acceptable Use Scale shared by Cari Wilson:

This is one way to embrace AI rather than fear and avoid it in classrooms. Again I ask:

How do we bolster creativity and productivity with AND without the use of Artificial Intelligence?

One way is to question the value of homework. Maybe it’s time to revisit our expectations of what is done at home. Give students work that bolsters creativity at home, and keep the real work of school at school. But whether or not homework is something that changes, what we do need to change is how we think about embracing AI in schools, and how we help students navigate it’s appropriate, effective, and even ethical use. If we don’t, then we really aren’t preparing our kids for today’s world, much less the future.

We aren’t going to AI proof schoolwork.

Top Risks 2024

1 Reply

I’d never heard of Eurasia Group before a good friend of mine, an investor, shared the infographic below with me yesterday. According to their website,

In 1998, Ian Bremmer founded Eurasia Group, the first firm devoted exclusively to helping investors and business decision-makers understand the impact of politics on the risks and opportunities in foreign markets. Ian’s idea—to bring political science to the investment community and to corporate decision-makers—launched an industry and positioned Eurasia Group to become the world leader in political risk analysis and consulting.

According to their ‘Top Risks 2024‘ report:

2024. Politically it’s the Voldemort of years. The annus horribilis. The year that must not be named.

Three wars will dominate world affairs: Russia vs. Ukraine, now in its third year; Israel vs. Hamas, now in its third month; and the United States vs. itself, ready to kick off at any moment.

Russia-Ukraine … is getting worse. Ukraine now stands to lose significant international interest and support. For the United States in particular, it’s become a distant second (and increasingly third or lower) policy priority. Despite hundreds of thousands of casualties, millions of displaced people, and a murderous hatred for the Russian regime shared by nearly every Ukrainian that will define the national identity of tens of millions for decades. Which is leading to more desperation on the part of the Ukrainian government, while Vladimir Putin’s Russia remains fully isolated from the West. The conflict is more likely to escalate, and Ukraine is on a path to being partitioned.

Israel-Hamas … is getting worse. There’s no obvious way to end the fighting, and whatever the military outcome, a dramatic increase in radicalization is guaranteed. Of Israeli Jews, feeling themselves globally isolated and even hated after facing the worst violence against them since the Holocaust. Of Palestinians, facing what they consider a genocide, with no opportunities for peace and no prospects of escape. Deep political divisions over the conflict run throughout the Middle East and across over one billion people in the broader Muslim world, not to mention in the United States and Europe.

And then there’s the biggest challenge in 2024 … the United States versus itself. Fully one-third of the global population will go to the polls this year, but an unprecedentedly dysfunctional US election will be by far the most consequential for the world’s security, stability, and economic outlook. The outcome will affect the fate of 8 billion people, and only 160 million Americans will have a say in it, with the winner to be decided by just tens of thousands of voters in a handful of swing states. The losing side—whether Democrats or Republicans—will consider the outcome illegitimate and be unprepared to accept it. The world’s most powerful country faces critical challenges to its core political institutions: free and fair elections, the peaceful transfer of power, and the checks and balances provided by the separation of powers. The political state of the union … is troubled indeed.

None of these three conflicts have adequate guardrails preventing them from getting worse. None have responsible leaders willing and able to fix, or at least clean up, the mess. Indeed, these leaders see their opponents (and their opponents’ supporters) as principal adversaries—“enemies of the people”—and are willing to use extralegal measures to ensure victory. Most problematically, none of the belligerents agree on what they’re fighting over.

Think about this, the Russia-Ukraine and the Israel-Hamas wars both take a back seat to the US election as the top risk of 2024. Both have no positive outcome in sight and they still don’t pose the same threat as a tight election result in the United States. I wish I could disagree, but I too see this as a genuine concern. What makes it worse is Risk #4 – Ungoverned AI, and specifically disinformation:

In a year when four billion people head to the polls, generative AI will be used by domestic and foreign actors—notably Russia—to influence electoral campaigns, stoke division, undermine trust in democracy, and sow political chaos on an unprecedented scale. Sharply divided Western societies, where voters increasingly access information from social media echo chambers, will be particularly vulnerable to manipulation. A crisis in global democracy is today more likely to be precipitated by AI-created and algorithm-driven disinformation than any other factor.

I want to explore the other risks as well, but by far my biggest concern for 2024 is the US election. My greatest fear is a close and contested election. The by-product of this would not just be tragic for the US, but for the entire world. I wish this was just hyperbole, but it’s not, and reading a report like this just magnifies concerns I already had. Buckle up, we are in for quite a ride in 2024.

You can get the full Top Risks 2024 white paper on their website, (or click the image below).

Conversational AI interface

Leave a reply

A future where we have conversations with an AI in order to have it do tasks for us is closer than you might think. Watch this clip, (full video here):

Imagine starting you day by asking an AI ‘assistant’, “What are the emails I need to look at today?” Then saying something like, “Respond to the first 2 for me, and draft an answer for the 3rd one for me to approve. And remind me what meetings I have today.” All while on the treadmill, or while shaving or even showering.

The AI calculates your calories used on your treadmill, and tracks what you eat, and even gives suggestions like, “You might want to add some protein to your meal, may I suggest eggs, you also have some tuna in the pantry, or a protein shake if you don’t want to make anything.”

Later you have the ever-present the AI in the room with you during a meeting and afterwards request of it, “Send us all a message with a summary of what we discussed, and include a ‘To Do’ list for each of us.”

Sitting for too long at work? The AI could suggest standing up, or using the stairs instead of the elevator. Hungry? Maybe your AI assistant recommends a snack because it read your sugar levels off of your watch’s health monitor, and it does this just as you are starting to feel hungry.

It could even remind you to call your mom, or do something kind for someone you love… and do so in a way that helps you feel good about it, not like it’s nagging you.

All this and a lot more without looking at a screen, or typing information into a laptop. This won’t be ready by the end of 2024, but it’s closer to 2024 than it is to 2030. This kind of futuristic engagement with a conversational AI is really just around the corner. And those ready to embrace it are really going to leave those who don’t behind, much like someone insisting on horse travel in an era of automobiles. Are you ready for the next level of AI?

What are you outsourcing?

Leave a reply

Alec Couros recently came to Coquitlam and gave a presentation on “The Promise and Challenges of Generative AI”. In this presentation he had a quote, “Outsource tasks, but not your thinking.”

I just googled it and found this LinkedIn post by Aodan Enright. (Worth reading but not directly connected to the use of AI.)

It’s incredible what is possible with AI… and it’s just getting better. People are starting businesses, writing books, creating new recipes, and in the case of students, writing essays and doing homework. I just saw a TikTok of a student who goes to their lecture and records it, runs it through AI to take out all the salient points, then has the AI tool create cue cards and test questions to help them study for upcoming tests. That’s pretty clever.

What’s also clever, but perhaps now wise, is having an AI tool write an essay for you, then running the essay through a paraphraser that breaks the AI structure of the essay so that it isn’t detectable by AI detectors. If you have the AI use the vocabulary of a high school student, and throw in a couple run-on sentences, then you’ve got an essay which not only AI detectors but teachers too would be hard pressed to accuse you of cheating. However, what have you learned?

This a worthy point to think about, and to discuss with students: How do you use AI to make your tasks easier, but not do the thinking for you? 

Because if you are using AI to do your thinking, you are essentially learning how to make yourself redundant in a world of ever-smarter AI. Don’t outsource your thinking… Keep your thinking cap on!

Is Generative AI “just a tool”?

Leave a reply

Dean Shareski asks,

Is Generative AI “just a tool”? When it comes to technology, the “just a tool” phrase has been used for decades. I understand the sentiment in that it suggests that as humans, we are in control and it simply responds to our lead. But I worry that this at times, diminishes the potential and also suggests that it’s neutral which many argue that technology is never neutral. I’d love to know what you think, is Generative AI “just a tool”?

My quick answer is ‘No’, because as Dean suggests tools have never been neutral.

Here are some things I’ve said before in ‘Transformative or just flashy educational tools?‘. This was 13 years ago, and I too used the term, ‘just a tool’:

“A tool is just a tool! I can use a hammer to build a house and I can use the same hammer on a human skull. It’s not the tool, but how you use it that matters.

A complimentary point: If I have a hammer and try to use it as a screwdriver, I won’t get much value from its’ use.”

My main message was, “A tool is just a tool! It’s not the tool, but how you use it that matters.”

I went on to ask, “What should we do with tools to make them great?” And I gave 6 suggestions:

  1. Give students choice
  2. Give students a voice
  3. Give students an audience
  4. Give students a place to collaborate
  5. Give students a place to lead
  6. Give students a digital place to learn

This evolved into a presentation, 7 Ways to Transform Your Classroom, but going back to Dean’s question, first of all I don’t think technology is ever neutral. I also think that technology need not be transformative yet always has the potential to be so if it really does advance capabilities beyond what was available before. So the problem is really about the term ‘just’ in ‘just a tool’. Nothing is just a tool. Generative AI is not “just a tool”.

We respond to technology, and technology transforms us. Just like Marshal McLuhan proposed that the communication medium itself, not the messages it carries, is what really matters, “The medium is the message”, the technology is the transformer, it alters us by our use of it, it is the medium that becomes the message.

Generative AI is going to transform a lot of both simple and complex tasks, a lot of jobs, and ultimately us.

What does this mean for schools, and for teaching and learning? There will be teachers and students trying to do old things in new ways, using AI like we used a thesaurus, a calculator, Wikipedia, and other tools to ‘enhance’ what we could do before. There will also be innovative teachers using AI to co-create assignments and to answer new questions, co-write essays, and co-produce images and movies.

There will be students using generative AI to do their work for them. There will be teachers fearing these tools, there will be teachers trying to police their use, and it’s going to get messy… That said, we will see more and more generative AI being the avenue through which students are themselves designing tasks and conceiving of new ways to use these tools.

And of course we will see innovative teachers using these tools in new ways…

How different is a marketing class when AI is helping students start a business and market a product with a $25 budget to actually get it going?

How different is a debate where generative AI is listening and assisting in response to the opposing team’s rebuttal?

How different are the special effects in an AI assisted movie? How much stronger is the AI assisted plot line? How much greater is the audience when the final product is shared online, with AI assisted search engine optimization?

Technology is not ‘just a tool’. We are going to respond to Generative AI and Generative AI is going to transform us. It is not just going to disrupt education, it’s going to disrupt our job market, our social media (it already has), and our daily lives. Many other tools, will not be just like the tools that came before them… changes will be accelerated, and the future is exciting.

‘Just a tool’ is deceiving and I have to agree with Dean that it underplays the possibilities of how these tools can be truly transformative. I’m not sure I would have said this 13 years ago, but I also didn’t see technology being as transformative back then. Generative AI isn’t a flash in the pan, it’s not something trendy that’s going to fade in popularity. No, it’s a tool that’s going to completely change landscapes… and ultimately us.

Alien perspective

Leave a reply

I think jokes like this are funny:

…because they hold a bit of truth.

We aren’t all that intelligent.

We draw imaginary lines on the globe to separate us. We fight wars in the name of angry Gods that are more concerned with our devotion than for peace and love. We care more about greed than about the environment. We spend more on weapons of destruction than we do on feeding the needy. We judge each other on superficial differences. We have unbelievable intellect, capable of incredible technological advancement, yet we let our monkey brains prevail.

Sure we exhibit some intelligence, we are intelligent viruses.

At least that’s what I think an objective alien visiting our planet would think.

Conversation on an alien ship observing earth:

“Give them another 100 years… if they figure out how to not kill each other and the planet, then let’s introduce ourselves.”

Right now I’m not terribly optimistic about what those aliens will find in our future? ‘Civilized’ humans? A desolate planet? Artificial intelligence treating us like we treat ‘unintelligent’ animals? Or more of the same bickering, posturing, warring, and separatist views of humans trying to usurp dominance over each other?

It would be funny if it wasn’t sad.

New learning paradigm

I heard something in a meeting recently that I haven’t heard in a while. It was in a meeting with some online educational leaders across the province and the topic of Chat GPT and AI came up. It’s really challenging in an online course, with limited opportunities for supervised work or tests, to know if a student is doing the work, or a parent or tutor, or Artificial Intelligence tools. That’s when a conversation came up that I’ve heard before. It was a bit of a stand on a soapbox diatribe, “If an assignment can be done by Chat GPT, then maybe the problem is in the assignment.”

That’s almost the exact line we started to hear about 15 years ago about Google… I might even have said it, “If you can just Google the answer to the question, then how good is the question?” Back then, this prompted some good discussions about assessment and what we valued in learning. But this is far more relevant to Google than it is to AI.

I can easily create a question that would be hard to Google. It is significantly harder to do the same with LLM’s – Large Language Models like Chat GPT. If I do a Google search I can’t find critical thinking challenges not already shared by someone else. However, I can ask Chat GPT to create answers to almost anything. Furthermore, I can ask it to create things like pro’s & con’s lists, then put those in point form, then do a rough draft of an essay, then improve on the essay. I can even ask it to use the vocabulary of a Grade 9 student. I can also give it a writing sample and ask it to write the essay in the same style.

LLM’s are not just a better Google, they are a paradigm shift. If we are trying to have conversations about how to catch cheaters, students using Chat GPT to do their work, we are stuck in the old paradigm. That said, I openly admit this is a much bigger problem in online learning where we don’t see and work closely with students in front of us. And we are heading into an era where there will be no way to verify what’s student work and what’s not, so it’s time to recognize the paradigm shift and start asking ourselves new questions…

The biggest questions we need to ask ourselves are how can we teach students to effectively use AI to help them learn, and what assignments can we create that ask them to use AI effectively to help them develop and share ideas and new learning?

Back when some teachers were saying, “Wikipedia is not a valid website to use as research and to cite.” Many more progressive educators were saying, “Wikipedia is a great place to start your research,” and, “Make sure you include the date you quoted the Wikipedia page because the page changes over time.” The new paradigm will see some teachers making students write essays in class on paper or on wifi-less internet-less computers, and other teachers will be sending students to Chat GPT and helping them understand how to write better prompts.

That’s the difference between old and new paradigm thinking and learning. The transition is going to be messy. Mistakes are going to be made, both by students and teachers. Where I’m excited is in thinking about how learning experiences are going to change. The thing about a paradigm shift is that it’s not just a slow transition but a leap into new territory. The learning experiences of the future will not be the same, and we can either try to hold on to the past, or we can get excited about the possibilities of the future.

AI and humans together

On Threads, Hank Green said, “AI isn’t scary because of what it’s going to do to humans, it’s scary because of what it’s going to allow humans to do to humans.”

I recently shared in, High versus low trust societies, examples of this with: more sophisticated scams; sensationalized click bait news titles and articles; and clever sales pitches, all ‘enhanced’ and improved by Artificial Intelligence. None of these are things AI is doing to us. All of them are ways AI can be used by people to take advantage of other people.

I quoted Hank’s Thread and said, “It’s just a tool, but so are guns, and look at how well we (miss)manage those!”

Overall I’m excited about how we will use AI to improve what we can do. There are already fields of medicine where AI can do thousands of hours of work in just a few hours. For example, drug discovery, “A multi-institutional team led by Harvard Medical School researchers has launched a platform that aims to optimize AI-driven drug discovery by developing more realistic data sets and higher-fidelity algorithms.”

The true power and potential of AI isn’t what AI can do on its own, it’s what humans and AI can do together.

But I also worry about people using amazing AI tools as weapons. For example, creating viruses or even dirty bombs. These are things that are out of reach for most people now, but AI might make such weapons both more affordable and more available… to anyone and everyone.

All this to say that Hank Green is right. “AI isn’t scary because of what it’s going to do to humans, it’s scary because of what it’s going to allow humans to do to humans.”

We are our own worst enemy.

The true danger and threat of AI isn’t what AI can do on its own, it’s what humans and AI can do together.