This Week in Edtech, 2/27/23
Introducing Insider Events & Speculations on the role of ChatGPT in Education
INTRODUCING… EDTECH INSIDERS EVENTS
At EdTech Insiders, we believe that our EdTech ecosystem thrives when we are connected as a community. From cross-pollinating ideas, to shared problem-solving, we are at our best when we are working together.
So we teamed up with our friends at the EdTech MBA Community, Cooley LLP, Owl Ventures, and Good Harbor Partners to host EdTech Summits in San Francisco and Boston this past month!
At these events, we had more than 300 edtech community members - entrepreneurs, investors, educators, and grad students - join us to participate in panel discussions, break-outs, networking, and small group dinners.
It was so inspiring to tap the collective wisdom and energy while also hearing from leaders like Matt Glotzbach from Quizlet and Jack Lynch from HMH.
As an extension of our podcast and newsletter, these community events provide a local touchstone for EdTech communities across the country and the globe.
Our next live event will be an EdTech SXSW Happy Hour on March 6 in Austin, TX, sponsored by EdTech Insiders and StartEd! We will have drinks, tacos, and the legendary Dan Carroll of Clever.
Join us at Seven Grand (405 E 7th St, Austin, TX 78701) on March 6 from 5:00 pm - 6:30pm CST.
Happy Hours are open, low-key events for our community to come together, and we hope to see you there! We will share more soon about upcoming happy hours:
LA (3/9)
SF (3/30)
San Diego (4/17)
London (6/21)
Philly (6/26)
NY (10/4)
TOP EDTECH HEADLINES
1. Department of Education Launches New Vendor Restriction Guidelines
As colleges and universities continue to work closely with online program management companies, the Department of Education is calling for transparency and regulation around OPMs and how they are contracted and compensated.
Beyond OPM providers, the potential redefinition of a whole swash of Edtech providers as Third-Party Servicers (TPS) could have huge effects on Edtech vendors of all kinds- although some observers (see Phil Hill)
2. Blowback Towards Big Tech Amidst Mental Health Crisis for Teen Girls
According to a new CDC report, teen girls are experiencing a mental health crisis, per the CDC, with elevated reports of sexual violence, anxiety and suicidal ideation.
Teen’s amount of time spent online and lack of feelings of connection at school are two impactful factors that are being explored.
Social media in particular is in the hotseat for regulation (I would expect the minimum age for social media accounts to be raised to 16 forthwith, with more to come)- what does this mean for Big Tech and EdTech moving forward?
3. ChatGPT and AI in Education
ChatGPT and other AI tools continue to polarize the education world; the first wave focused on the division between plagiarism fears (Techcrunch this week showed that the current crop ani-plagiarism tools are lacking) and the predictable bullishness of Venture Capitalists.
The next wave is more nuanced, with new fears for teens’ already fragile mental health (Inside Higher Ed) and new champions of AI as a teachable moment for educators (Times Higher Ed).
A great quote for the latter per Leanne Ramer from Simon Fraser University:
“Across institutions of “higher” learning, the collective response to ChatGPT is being hindered by a misplaced focus on implications for academic integrity. Can students use AI tools to “cheat”? Maybe. Should we be assessing university and college students on work so trivial that a beta-version large language model can readily earn a high grade? No.”
Either way, we can expect a slew of new AI-based Edtech companies this year.
4. How The Decrease in Edtech Funding Will Impact Schools
Funding for EdTech companies has dropped significantly since 2021; what does this mean for schools and how will EdTech products be integrated into schools moving forward?
LISTEN
Recent Episodes of the Edtech Insiders Podcast »
AI Superstar Kian Katanforoosh, CEO of Workera, on digital upskilling
A Rare Peek into the Chinese Edtech ecosystem from Andrew Shewbart of ALO7
Building the One-Stop Shop for Edtech Enterprise Sales with Siya Raj Purohit of Amazon Marketplace
THE BIG READ: THE FUTURE OF AI
Assuming you haven’t been in a sensory deprivation chamber for the past few months, you have heard a lot about ChatGPT; it’s been banned, celebrated, has passed medical, business and law school exams, and is now falling in love with journalists.
While some in Education fear ChatGPT, we at Edtech Insiders are taking a longer view, considering ChatGPT as a paradigm shift in education and work.
With that in mind, we collaborated with Edtech CEO and AI maven Hanna Celina of Kinnu.xyz to create a speculative science fiction narrative:
Beyond imagination: The Future of AI
It’s been four decades since the pivotal year 2022, when OpenAI launched its first groundbreaking ChatGPT AI system. As we all know, the world is now powered by fusion, and we successfully halted and reversed global warming; recently, we’ve made major progress in bringing back extinct species. The World Government is thriving; wellness, lifespan and wealth for all our global citizens continues to grow. But lest we forget: it was a close call - the story of the Golden Age of Humanity almost didn’t have a happy ending.
Madelaine gazed over the long marshland grasses, which waved gently in the wind, and into the sunny horizon. She leaned absent-mindedly toward the sun, leaning over the railing of her balcony.
Mina, playing in the tall grass, saw her grandmother lost in thought and, after a brief hesitation, called out - “Mammi, can I ask you a question?”. Madelaine looked down towards her and smiled, and Mina continued: “How did you decide to be an interstellar ecosystem designer?”.
Madelaine smiled gently. “Actually, I didn’t”, she said, looking down at the thirteen year old. “In fact, when I first started working, such a job didn’t even exist. I was what used to be called a writer.”
Mina looked at her, confused.
“What’s ‘a writer’?”
Madelaine sighed and waved Mina inside, - “Want to go on a little field trip through history?”
Chapter 1: Humans in the driver’s seat
Madelaine gestured for Mina to sit down, by the comfortable armchair in the living room, pointing knowingly to her temple as she took a seat on an ottoman.
“Let’s think back to 2025, way before you were born. At that time, AI was only an interface that people used to speed up creative projects or to make predictions and classifications. You would use a keyboard or a voice command to generate words - we called them ‘prompts’ - which the AI or neural network turned into what it saw as the most probable output. People were still the ones deciding what to ask about - they were in the driver’s seat.”
Mina raised her eyebrows in surprise.
“Imagine a woman parking her car and walking into an office. She sits by the desk with a calendar pointing to Christmas, and opens her notebook filled with handwritten notes and types something into a computer. That was me.”
“Before these first generative AIs, I was a writer – that means I would write my own stories and articles down without any AI at all!”
“We still used letters back in those days, which I know seems like such an arcane technology - communication without Imagodai seems so limited. It seems strange to explain and write down everything instead of just converting thoughts to images and models.” Madelaine smirked and continued.
“Hm.” Mina grunted, looking over the horizon and lost in thought. Such a world felt unbelievably primitive, almost unimaginable to her, Madelaine knew.
“Then the first AI tools started coming out, and we were amazed at them. Back then, if you can imagine it, you could only make images — machine dreams, they were called — or ask for written pieces of information, and soon after that, videos and simple 3D simulations. It was still a couple of years later that the AI would write whole books, design games and make movies from a single one-sentence prompt. It worked a little bit like your ‘Makebot’ toy; but back then, that was all we could do.”
“But to us, these AIs meant a whole new world of possibility. It felt like the keyboard was suddenly a magic wand - you just needed to point AI in the right direction, pay a little fee for the computer power and *puff* - magically your answer appeared. Sometimes these answers were simplistic, biased or even factually incorrect — after all, these early neural networks were only the average of everything that humanity, in all its endless bias and confusion, had ever created. But it was still exciting – our first glimpse of the world-changing developments that were to come.”
Madalaine chuckled: “At first, the only thing some teachers thought about was whether students would use the AI to ‘cheat’ on their homework and tests.”
“What’s ‘cheating’?’ Asked Mina.
“It’s hard to explain. Back then, students were asked to write their own essays, or answer questions, just relying on their own memories and brains, without any AI assistance. So if a student did use AI, it would be… against the rules.”
“Whooooooaaaa….” Mina gazed into the middle distance, trying to imagine this state of affairs.
“Luckily, it didn’t take long before the world realized that working alongside AI was actually one of the core goals of education. They started calling it “Augmented Intelligence” and teaching it in schools. It was an amazing time.”
She continued: “Back then, the models that drove these AI generators were called ‘large language models’ or ‘LLMs’ for short, and they helped humanity automate and scale ‘small c’ creative tasks - those which required a relatively light level of thought and imagination.”
“Small c?”
“That’s what they called it. Maybe this is one way to think about it: if ‘small c’ creativity was about doing creative work – such as a simple poem or a drawing or making music – ‘big c’ creativity was about having original and insightful ideas that nobody had ever had before. We knew that the greatest works of human-created art were not just well-produced – they also had incredible ideas behind them. Back then, humans thought we could hold on to ‘big C’ creativity, as a skillset that was safe from AI replacement.”
“But as you know… we were wrong. There were fewer and fewer professional writers by the year; and it was 2028 when I started my first job as a designer. I got a job creating AI models with feedback loops - that was a process in which one initial question or prompt would be fed back into the neural network to create a cascade of verbal, visual and video output, which was then calibrated and improved over and over again with the help of us humans giving feedback. I did this for a publishing house, and we produced books for school kids throughout the 2030’s. People loved the models I trained – some of them are still popular to this day, and I was told I was very good at my job.
“Wait, what does it mean that you were good at creating models? Weren’t all of the AI models perfectly personalised to each human- ‘in the eye of the beholder’?” asked Mina, confused.
Madalaine smiled. “Not yet. It was still early days of the technology and the models were training themselves on human generated data - some of their creations were brilliant… but some of them were not even worth their own computing power. We would tell the model which ideas were better than others or more ‘creative’ and interesting and fun from a human perspective; there was a whole team dedicated to correcting factual errors, and another team dedicated to catching biases that were left over from our ancestors’ silly ideas about how people were different from one another. That’s a whole other story. Between us, we could produce a full kids novel in about two weeks, which, if you can believe it, was much faster than human writers. All of our teams were proud of our work; we used to say that we were ‘in the loop’.
“After a while, it became harder to know how to really guide the AI, because AI started excelling way past people at most tasks that would have previously been labelled as ‘creative’. The job was still fun, because the AI was coming up with amazing stories; we’d all be on the edge of our seats waiting for the next chapter. Gradually, the fact-checking and ethics teams got smaller, as the AI became less error-prone, but it still made plenty of mistakes.”
“It’s hard to imagine an AI making a… mistake.” Mina said, dreamily.
“Back then it was normal! In fact, we thought that humans had the advantage. The kinds of things we thought were: How can AI understand what it means to be human, or to have a consciousness? Does empathy boil down to a specialised kind of probability or is it a uniquely human quality? Can AI, which is basically a monkey with an infinite typewriter, generate new art which touches humans in the same way as the great human artists of the past - Van Gogh, Shakespeare, Liszt, Beyoncé? At the time, AI still relied on the human experience as its only filter for what was ‘bittersweet’ or ‘beautiful’ or ‘heartbreaking’. We had no idea what was coming next.”
Stay tuned for upcoming chapters…
FUNDING ROUNDS
US
India
Europe
Guider Announces the Upcoming Launch of Their Category-Defining Educational App for Muslims
German edtech Knowunity, a TikTok for schoolwork, raises €9m Series A extension
EdTech Bloomcoding Grabs €1M Funding to Push Coding with VR in CEE
Thanks for reading Edtech Insiders! Subscribe for free to receive weekly updates and support the podcast.