Why is Sal Khan an Optimist?

In his Vancouver, BC Ted Talk on AI, Sal Khan describes the powerful potential for AI to transform education in the near term.  

Kahn offers exemplars of the intuitive AI super tutor brilliantly accelerating student performance, not just 1:1, but in a scalable way.  He shows how students will have access to life coaches, educational coaches, and writing coaches to support the academic journey of a child.  These AI offerings stand to increase confidence and improve outcomes.

Khan also demonstrates how AI uses traditional ‘teacher tricks’ like wait time, Socratic methodologies, answering questions with questions, and the classic “How did you get your answer?” He offers teachers hopeful ideas about personalizing learning and efficient lesson planning on complex issues.

Use AI to enhance HI – human potential, human intelligence, and human purpose.

What does it mean that “Bob doesn’t know how it works?”

I’ve lived through several massive shifts in technology over my lifetime:  personal computer, internet, smart phone.  These shifts have remarkably changed my daily lived experience.  I have always valued the people in my world who were experts on these technologies.  They understood the inner workings, had language for how it worked, and researched the solutions for security and reliability.  These experts were treasures.

Now, with AI, the experts are telling me they don’t know how it works.  They don’t understand and can’t validate the outputs.  The interpretability methods are still emerging because of the enormous complexity of the processes of AI.

While researchers are working on interpretability methods, even the MIT researchers say they all should be taken with a ‘grain of salt.’  We just aren’t sure how it works yet.  

So, students need to be aware of these weaknesses and educated on the ethical implications.  We need to remain critical thinkers and ethical decision-makers as we step into this realm.

How is AI a mirror?  What does it show us?

All of our human foibles, biases, and misinformation are embedded into AI.  It is simply a mirror of our collective.  The same racism, coercion, and falsehoods are likely.  Fundamentally, its aim to sound, respond, and communicate like a human.  

In other words, AI is non-neutral.  The biases, culture, and perspectives of its creators are deeply embedded into the algorithms of AI.  

UNESCO offered an example last month in their article on the ethical dilemmas of AI.  They invited us to type the prompt “greatest leaders of all time” into a favorite search engine and check out the list of prominent male personalities.  How many women are listed?

These gender biases are just one example of how AI is non-neutral, a mirror of our current society.

When I asked ChatGPT how it was like a mirror to humanity.  It gave me the following three reasons:

• Reflecting Knowledge
• Reflecting Biases
• Mirroring Human Behavior

Sounds about right to me.  Embedded into the response were a couple of disclaimers:

• It’s important to be aware of this aspect and critically analyze the information provided by ChatGPT, just as one would critically evaluate information obtained from different human sources.
• ChatGPT is an AI system and not a true reflection of human consciousness or emotions.

How do school leaders need to ensure we are teaching the discernment and critical thinking required to use employ these as a mirror, not a master?

What is uniquely human? How can we amplify those uniquely human traits? 

Over millenia, we’ve used metaphor to understand the mysteries of life.  Consider energy as a wave or a particle, God as a mother hen, or your current project as an open file.  

Similarly, cognitive science purports that our minds function like a computer.  In other words, our mental states and cognitive processes are mechanistic.  We use vocabulary like record, retrieve, file, and compute to describe the function of our minds.  The theory doesn’t include our bodies.

Are these metaphors newly unworthy?  Are we now grappling with how we’ve minimized our own beings to be too mechanistic and computer-ish over the last couple of generations?

I believe school leaders need to actively explore with students the uniquely human traits, virtues, ethics, common sense, and souls of humanity.  What is the difference between a sentient artificial intelligence and me?

What is our roadmap for the ethical and practical considerations of AI in schools?

Without regulatory frameworks or ethical tools, we must proceed with caution.  From my perspective, it is definitely a misstep to ban these tools.  And, it is a misstep to proceed without deep inquiry into how to develop our practices and approaches with wisdom.  

As a starting point, as Khan advocates, our use of AI tools needs to amplify positive use cases and develop guardrails for competency-based learning.  Critical and creative thinking must be overlaid with all artificial intelligence tools.  Additionally, assessment needs to be formative and authentic.  In other words, how can we use these tools to deliver powerful learning experiences?

Most importantly, our human relationships need to be preserved, nurtured, and upheld as foundational for the unfolding of extraordinary human potential.

For the sake of the children,

Karine