Do you want be 40% more productive at work? Do you want to do higher quality work? Do you want to learn more? Do you want to get a job or a promotion? Use generative AI
Imagine that all the employers in your area knew graduates of your institution had generative AI skills.
There are a couple terms that scary: “artificial intelligence” and “cyborgs.”
A lot of people are worried about AI taking over the world, or at least taking their job. Others worry about becoming cyborgs: half silicon individuals with computer chips planted in their brains. They worry this is what we mean when we talk about “augmenting” intelligence
But that’s not really what it means.
What it means is that is that if you use a basic AI tool (a generative AI tool in most instances) to help you do your work or learn you’ll augment your intelligence and you will perform better. This is no longer a controversial claim.
I use AI tools every day to augment my intelligence. I use(d) them to --
Write, edit, rewrite, strengthen my expression for use in different contexts
Design recruiting videos for my debate team (this would have taken me at least a day to do before generative AI, it took me 2 minutes)
(Re)organize literature reviews
Translate a 1-minute video to Spanish that I recored in English to help recruit some students new to the US to the debate team
Translate a 1-minute video about a debate class to Chinese-speaking parents
Generate a list of vocabulary words related to a debate topic for my high school students. I even asked it to include an example related to the topic
Translate debate materials for Spanish students
Identify trends in participation across debate tournaments using Code Interpreter (now Advanced Data Analytics)
Write drafts of emails
Produce some photos for some blog posts and notes that I hope piqued interest
For a list of ideas to manage a situation
For a nicer way to say things
To explain things to people in ways that will help them understand them
To provide a historical reference I can’t remember the details ofBuild a test bot for a school district. That took me an hour and required no tech skills.
Capture students’ attention. It was late in the afternoon and I was losing the students’ attention in class the other day. I pulled out my phone and started talking to Pi about our class topic. That got their attention.
These things all help me with output.
It also helps me learn quickly and in ways a simple internet search cannot –
It helps me understand things that are complicated and relatively new to me (back propagation, parameters). Parameters wasn’t too complicated, but back propagation took me a bit, so I asked it to explain it to me like I was 10, 15, in college, then in college with a few computer classes, etc. I directed it to teach me and it did a good job.
Explain the history of an idea, adapted to my education level in the area
See the relationships between some concepts I was working on by displaying it in a table
Answers my questions any time I have the desire and availability to learn
These are just some simple things for every-day work. When others use it, “on the job,” how does it go?
Pretty well. In a recent study conducted of some Boston Consulting Group employees (7% of their consultants), they killed it. As summarized by Wharton School professor Ethan Mollick:
(F)or 18 different tasks selected to be realistic samples of the kinds of work done at an elite consulting company, consultants using ChatGPT-4 outperformed those who did not, by a lot. On every dimension. Every way we measured performance. Consultants using AI finished 12.2% more tasks on average, completed tasks 25.1% more quickly, and produced 40% higher quality results than those without. Those are some very big impacts. Now, let’s add in the nuance. First, it is important to know that this effort was multidisciplinary, involving multiple types of experiments and hundreds of interviews, conducted by a great team, including the Harvard social scientists Fabrizio Dell’Acqua, Edward McFowland III, and Karim Lakhani; Hila Lifshitz-Assaf from Warwick Business School and Katherine Kellogg of MIT (plus myself). Saran Rajendran, Lisa Krayer, and François Candelon ran the experiment on the BCG side, using a full 7% of its consulting force (758 consultants)…. There were creative tasks (“Propose at least 10 ideas for a new shoe targeting an underserved market or sport.”), analytical tasks (“Segment the footwear industry market based on users.”), writing and marketing taskfor s (“Draft a press release marketing copy for your product.”), and persuasiveness tasks (“Pen an inspirational memo to employees detailing why your product would outshine competitors.”). We even checked with a shoe company executive to ensure that this work was realistic - they were. And, knowing AI, these are tasks that we might expect to be inside the frontiline with our tSOme conheories, and as we have discussed, we found that the consultants with AI access did significantly better, whether we briefly introduced them to AI first (the “overview” group in the diagram) or did not. This was true for every measurement, whether the time it took to complete tasks, the number of tasks completed overall (we gave them an overall time limit) or the quality of the outputs… The consultants who scored the worst when we assessed them at the start of the experiment had the biggest jump in their performance, 43%, when they got to use AI. The top consultants still got a boost, but less of one. Looking at these results, I do not think enough people are considering what it means when a technology raises all workers to the top tiers of performance.
This isn’t an isolated study. See also Noy & Zhang. And at this point it’s becoming common sense. Use it for a few hours and you’ll agree with me.
[Side note: I ran into a friend who I hadn’t seen in a while at a party and we ended up talking about ChatGPT. He runs at research center at a major university. He sent me an email the other day: “The morning after I saw you, I bought my OpenAI subscription so that I can access ChatGPT 4. I agree with you - it's fascinating and very useful. For my job, it's like having a research assistant at my beck and call. Yes, I have to check the validity of what ChatGPT spits out. But that's true for a human research assistant as well!”] This reminds me of the Ethan Mollock’s post on interns. My friend is smart like professor Mollick.
And think about this: This is with an AI that is, relative to what’s coming, very rudimentary. What productivity gains will we have in a year as people learn more about how to use the tools? As the tools improve?
It will be big and it will transform work (it already is).
How much will it transform work?
Earlier this year, Goldman Sachs estimated that 63% of work will be complemented by AI. McKinsey notes, “The biggest impact for knowledge workers that we can state with certainty is that generative AI is likely to significantly change their mix of work activities.… To capture the full benefits of generative AI to make knowledge work more productive, employers, policy makers, and broader ecosystems would need to establish clear guidelines and guardrails—and workers would need to see these tools not as job destroyers but as work enhancers.”
Of course, this can mean slightly different things. People could end up working with AI or AI could end up doing part of their job, leaving them to do other things we may need humans for. It will likely be some combination, but it’s hard to imagine too many humans who are in this 63% not needing to understand how AI works and how to work with it.
What does this mean for educators? Most of my blog readers are educators, so I’ll take a stab.
(1) Your employer should pay for ChatGPT4 (or a similar tool). Assuming you are an education administrator, you are likely making 100-200K/year (or more). If you finish 10% more tasks, get 25% done faster, and get 40% done better, then your employer will get that out of you for a 1-2% raise ($20/month) for a .0024% raise (or less).
This is my friend Carl. He’s the CTO of a county education leadership team and we recommended that his team get ChatGPT4. He agreed and now he’s waiting for OpenAI to approve their ChatGPT4 enterprise account. We are so aligned we wore the same thing to the AI training day.
(2) Learn how to use it and train other people to use it.
(3) Think about who you want to be and what you want your school to be known for. Do you want to be the educational leader who can guarantee to the community that no one in your high school or university who graduates will have ever used generative AI and been trained to do so, or do you want your graduates to be people local employers know have never used generative AI?
(4) Help/enable your students to use the tools. Teach them the best you can (we are all learning).
After my presentation at the Cottesmore Conference in May, someone asked where to get started. I said the best way to start was to just try out the tools. I was happy when Sam Altman suggested the same to Minerva founder Ben Nelson a few days ago when Nelson asked him for suggestions as to how schools should get started with AI.
This isn’t complicated. We are transitioning to an AI World. Our graduates are going to live in it and work in it (last year’s graduates already are). Those who live in an AI World and know how to work with AI will be more employable, as an ability to work with AI one of the most in-demand skills.
Are there downsides? Of course. In the work world, people may rely on AI too much and “fall asleep at the wheel.” Those who aren’t trained will use it poorly.
Students who are receiving assignment that have not been adapted to the word of AI might “cheat” if you don’t block it from their school computers, but every kid who can afford more than a school computer and wants to “cheat” is going to “cheat” anyhow. Generative AI is everywhere and there is no way to control it.
And are we really going to hold the future of our educational system and opportunities to prepare students for the AI world hostage to the limited number of students who don’t want to learn anything and are going to try to find every possible way to cheat? Motivated students can learn anything they want 24 hours a day 7 days a week, as they’ll literally be able to have multimodal conversations with all the world’s knowledge any time they want. In the end, Students who have equitable access to AI, are taught how to use it well will learn the most, and are motivated to learn will skyrocket over their peers.
These students will have amazing tutors and be prepared for the world of changing jobs.
This is a clip from an interview with Dr. Ilya Sutskever, the Chief Scientist at OpenAI who lead some of the initial deep learning discoveries with Geoffrey Hinton. Having started taking university in the 8th grade, he's has above average intelligence. In it, he talks about the development of AI tutors and the importance of students using the tools.
Joyce Mullen, the President and CEO of Insight Enterprises, notes, “the people who learn to use the tools most effectively likely will replace those who don’t.”
Here is a very basic example of how even current technology can combine vision and text to support tutoring.
How long do you think it will be before students can ask it or a similar application to demo a science experiment?
It’s time for our administrators, faculty, staff, and students to start using AI. Failing to do so is like leaving your second brain at home when you go to school while the world races ahead.