Could AI Help Us, Not Replace Us?  🔗 **Source:** [Could AI Help Us, Not Replace Us?](https://share.snipd.com/episode/3369455e-8f32-4167-a6b2-f93274a48dfb) ## Highlights - **Humanistic AI Is Augmentation Not Replacement** - Tom Gruber defines humanistic AI as AI designed to augment and collaborate with humans rather than replace them. - He contrasts a machine-centric path that automates work with a human-centric path that amplifies human abilities and solves big problems. Transcript: Tom Gruber I'm here to offer you a new way to think about my field, artificial intelligence. I think the purpose of AI is to empower humans with machine intelligence. As machines get smarter, we get smarter. I call this humanistic AI, artificial intelligence designed to meet human needs by collaborating and augmenting people. Manoush Zomorodi That was around the time that you gave your TED Talk in which you introduced the idea of humanistic AI. Can you define what that was in your mind? Tom Gruber The idea was really like, where do we stand in the world of the future of AI? And so there was really two paths I saw. One path was, you can sort of say machine intelligence or machine-centric, delighting and celebrating how smart machines were getting relative to humans. And there were businesses being formed to use that to automate human work. And then the other way of looking at it, which I call humanistic AI, which wasn't just my idea, It was a lot of people had this idea, was that the purpose of AI should be to actually help people Do things they're trying to do by either augmenting their intelligence or collaborating with them as an intelligence. And so it turns out that down the road now, that was a watershed because now we see companies that are raising money by the billions with the explicit goal of automating white-collar Work. And then we have other companies that are raising billions of dollars and saying that our job is to help solve some of humanity's big problems and make people smarter. So that's why I wanted to give it a label, and that's why I call it humanistic AI. ([Time 0:04:26](https://share.snipd.com/snip/96df2f7a-09d4-4202-a70e-c7bbeb5202f8)) - **Siri Cofounder Built Assistive Voice Tech First** - Tom Gruber built an early communications assistant for people who couldn't speak using a tiny language model to predict words from a single muscle switch. - The system produced voice output (primitive synth) enabling sentence communication for users with cerebral palsy and ALS. Transcript: Tom Gruber The models that we could build then were nothing like human intelligence. They were just crude approximations. So then we decided, OK, well, how would you augment human intelligence? And one of the ways is to shore up where there's some kind of impairment in human capability. And so the first project I did was to build what I call the communications assistant for people who can't speak, who have like cerebral palsy in this case, but also ALS and folks like that, That have neurological conditions that makes it hard to speak like we're speaking here. And so we built an AI program that had an LLM, a little language model, maybe a TLM, tiny language model, that actually predicted the next word and the next phrase based on single motor Action, like a switch on their muscle. They could actually communicate in sentences. ([Time 0:08:36](https://share.snipd.com/snip/7bd905a8-1269-4df0-ab29-0bb61de7aae3)) - **Optimize AI For Human Benefit Not Profit** - Design AI objective functions to optimize for measurable human benefit rather than profit alone. - That requires research into detecting harms, building theories of human benefit, and engineering guardrails into models. Transcript: Tom Gruber Yeah, the humanistic AI framework would say that the objective function of the AI, that is the thing that the AI is optimizing for, should be human benefit, not, say, profit or something Else. And human benefit is hard to measure. So there's no easy prescription. So we have to decide if we're going to put guardrails on a thing, we have to solve the engineering and scientific problems of detecting that something the AI is doing is harmful to humans, We have to build theories of human harm and human benefit into our objective functions. So that's what I think we should be working on. ([Time 0:12:00](https://share.snipd.com/snip/542ef3a9-e939-48b7-8e5b-e0055839ad4a)) - **Apps Will Compete To Make Models Humanistic** - Foundation models are costly and rare to build, but applications can compete on human-centered design atop them. - This mirrors the app ecosystem where many apps can choose to prioritize safety or human benefit. Transcript: Manoush Zomorodi That kind of how Apple created the app ecosystem and you have to adhere to certain privacy standards? Tom Gruber Yeah, that's right. Well, that's a very good analogy there. It would be great if there was something like an app store for AI that would do the work of at least minimally establishing whether an application is in the benefit of humans or not. ([Time 0:18:33](https://share.snipd.com/snip/6fbc0c6e-d295-40a3-8bff-fea8bc83899b)) - **Build AI That Acts Like A Protective Caregiver** - Treat AI like a protective caregiver—collect deep personal data only if systems are aligned with users' interests. - Tom Gruber calls this Big Mother: use data to nurture and teach, not to manipulate attention or maximize profit. Transcript: Tom Gruber I tried to figure out what would be the right symbol. And I found, I think, an African elephant is kind of a cool big mother. It's smart, protective. The matriarchs are the ones that run the herds. They're the wisest and smartest and so on. So imagine a female African elephant with her calf. And think about her value alignment. She will do anything to protect the calf. And obviously humans too, human mothers. They nurture their children. They teach them right from wrong and truth from falsity. And then they show them skills to survive in the environment they're in. That's what AIs can and should do. Machines can now have access to everything in your life that's digital. And with that, you can build amazingly good recommendations about how to use your attention. Maybe give you insights on things you could discover or learn from. At the same time, that same data is extremely powerful if it's used in a surveillance economy against your interests to addict you, to make you buy things. This is why it's a real ethical choice. How you optimize the AI is actually a massively important societal choice today. It's not a technical thing that we leave to the engineers or to the boards of directors of companies. It has to be done at the society level. And so I think Big Mother is the way to say, go ahead, have a lot of data. Like, mothers know everything about their kids, but mothers are aligned with their kids' interests. ([Time 0:19:27](https://share.snipd.com/snip/eab90b20-82e4-4730-acdb-c998ef59fe19)) - **Education AI Must Combine Neuroscience With Personalization** - Priya Lakhani integrates neuroscience and learning science with AI to personalize education for every brain. - Century Tech targets teacher workload reduction plus tailored learning rather than simple recommender-style personalization. Transcript: Manoush Zomorodi Century Tech is the educational tech company that Priya founded in 2013. She says the platform's goal is to support overworked teachers, not take their place. There is, of course, a big debate in education right now about the role of AI in teaching. Some believe funding should go straight into schools, not into technology that's trying to make money off of learning. Others think the only way to make education accessible to every child will be with the help of AI. That's the camp that Priya is in. But she says that doesn't mean replacing the hard, even grueling work that happens in the classroom between teachers and their students. Here's Priya Lakhani on the TED stage. Priya Lakhani We need to combine artificial intelligence with neuroscientific theory and the learning sciences to learn how every single brain in this room learns. ([Time 0:24:25](https://share.snipd.com/snip/45a67cda-b055-4b25-aeac-caccfacfe7f3)) - **Use AI To Empower Teachers With Targeted Insights** - Use AI to surface real-time insights so teachers can make targeted interventions rather than replace classroom instruction. - Century Tech shows dashboards highlighting misconceptions so teachers can do focused, small-group teaching. Transcript: Priya Lakhani They could either set them learning material. Usually they set them a set of questions. The teacher has their dashboard open and they can see how the students are performing. So information would light up about who is struggling. The teacher would then walk around the classroom and be able to make those interventions real time. Okay, so this student, why did you answer that question in that way? And the teacher's then using their expertise. So that's a blended learning environment. And the teacher would then utilise that information to then be able to not just stand in front of the class assuming everyone's taking in the knowledge and asking people to put their Hands up and spot-checking with their knowledge. You can walk around and have that very sort of targeted intervention with students. Okay, so that's in the classroom. But what about at home? So half of them will use it as teacher sets homework. To Manoush, you've been given this assignment by your teacher. It's on Pythagoras' theorem. You may be a student where before that it has automatically given you some maths work on roots and powers because it knows that you don't understand that. And there's no point you're doing the Pythagoras work if you don't understand roots and powers. In the same way, right, if you master that work, it can then do what's called a smart recommendation and stretch you as well. So some people will use it for assignments in that way. And then the others will use it in a flipped learning way. So teachers will be planning on teaching a lesson. They'll say, right, we would like you to learn the lesson on Century the week before. Students do that in their own time. Teachers then receive the information of how the students have done. And then that lesson the following week is a far more targeted lesson as to what did they understand, what didn't understand. So 50% of the platform is about the teacher, because we fundamentally believe that empowering teachers is one of the best ways to improve education. ([Time 0:26:30](https://share.snipd.com/snip/df1a250b-cfc5-4b5f-b093-ffd576bf65b4)) - **Fluent AI Output Can Mask Shallow Learning** - Durable learning requires productive struggle; fluency from AI feels like learning but often isn't retained. - Priya argues AI should force generation of answers and provide structured feedback against teacher-designed rubrics. Transcript: Priya Lakhani Think about how we felt when we first used ChatGPT. I think we thought, wow, I never need to do any work ever again. This is amazing. And then it hallucinated. And I think we've ended up with this sort of sinking realization of acceptance, right? That the shortcuts don't really replace the work. They're very helpful, but we still need to learn and we need to think. Now, when we read those long answers that an LLM chatbot gives us, it feels very fluent. The problem is, is that fluency we often mistake for learning. What we actually know about learning is that learning requires what researchers called a productive struggle. It's this sort of mental effort that builds understanding. Sustained mental effort strengthens the parts of the brain and it's positively correlated with growth in the brain. Durable learning does not come from shortcuts. It comes from certain types of effort. And this is why AI is amazing for education, because AI can spot patterns in how we all learn. It can force you to generate an answer rather than just reveal the answer. ([Time 0:33:31](https://share.snipd.com/snip/c083c4b6-0982-48c6-9482-cfa9931600ef)) - **Measure EdTech Success By Time Off Screen** - Measure AI success by how quickly students can leave the screen, not engagement metrics. - Priya sets a guideline of about 1.5 hours per week on the platform and warns excessive use signals poor performance. Transcript: Manoush Zomorodi You measuring for then? If not engagement, then what? Priya Lakhani My key success metric is how quickly can I get this kid off the screen? And so we give guidance that's completely the opposite of a typical company. We say to schools, you know what, you really should not be on our system for more than about an hour and a half a week. If you're sitting there for nine hours a week, this system is actually not performing at the level in which we would hope it to be. ([Time 0:36:36](https://share.snipd.com/snip/50fec6ec-384c-4223-992a-0165a44936d6)) - **Job Disruption Has Precedent And New Jobs Will Emerge** - Vlad Tenev frames AI-driven job disruption as historical continuity: past eras lost jobs but created many new ones. - He argues AI will create new roles and opportunities, and humanity has repeatedly adapted to technological shifts. Transcript: Vlad Tenev Take a moment and reflect back upon our lives when we were 20 years old. Think about the opportunities for work and career that lay in front of you. How many of you had a pretty good idea of what you wanted to do for your career? Not too many. How many were overwhelmed by all the options? I know I felt the same. Manoush Zomorodi Vlad recently gave a talk called AI is Coming for Your Job, Now What? In it, he told his own story of being 20 years old and graduating from Stanford University with a degree in mathematics. Vlad Tenev Nobody had sat me down to tell me that my pure math major wasn't going to be the most desirable qualification, and I probably wouldn't have listened if they did. Now, my first month in graduate school, Lehman Brothers went under, the start of the global financial crisis. Most of my friends, particularly the ones that felt the most secure, found themselves packing up their cubicles. Some of us wondered whether the economy would recover at all or whether we were in store for another decade-long Great Depression. But amidst the uncertainty, some of us found a source of optimism. The iPhone app store came out that very same year, 2008. Manoush Zomorodi And the idea that anyone could build a digital game or service that could be delivered to millions of people and a device that lived in their pockets? Well, it was the digital equivalent of when the first pioneers found gold in California in the 1840s. Vlad was all in. Vlad Tenev I still remember when the instruction manual for how to build iPhone apps was released. I was up all night reading it, learning, trying to understand. I saw an opportunity for a new level playing field. Pretty much everything I've done since then was a product both of the economic malaise but the technological optimism of the time. But times have changed. The average 20-year today also has quite a bit of fear. But this time, emerging technology is not the antidote to that fear. It's the source. And they're asking themselves, will that career I'm looking at even be around in 10 years? One reason why it feels different this time is because AI, unlike the iPhone, is the first tool that we've built that's capable of leaving the toolbox, and we don't yet know its limits. A few years ago, I founded another company with the mission to build mathematical superintelligence, artificial intelligence that can reason and solve problems better than any Mathematician. I always thought of mathematics as the pinnacle of human intellectual activity. So a superhuman AI at mathematics could potentially be superhuman at everything. Combine that with my day job, which is running a global financial services platform, it's led to me spending a lot of time pondering one very important question. What do we do in a world where the vast majority of today's jobs are gone? And I want to analyze this question rationally, without fear and hyperbole. One way to do it is to look back through history and see if there's been a time where we faced this type of job disruption before at anything near these levels and how we as humans have navigated It. Now, I'm a technologist, not a historian. So with that caveat, let's go back in time to a world a 20-year would have known tens of thousands of years ago. Manoush Zomorodi In approximately 50,000 BCE, most people were hunters, gatherers, or toolmakers. Very few of us today know anyone with those job descriptions. Vlad Tenev The main occupations of the Paleolithic era are largely gone, but they didn't disappear overnight. Instead, they were subdivided into lots of other more specialized jobs. Manoush Zomorodi The next era, the Neolithic era, saw all kinds of different vocations popping up, thanks to advances in how people stored their food and built shelter, domesticated plants and animals. Vlad Tenev Humans have mastered a few new things, farming, keeping livestock. The invention of these things allowed us to spend more time doing what we consider creative work and less time on pure survival and subsistence. And this opened up a lot of new jobs. You had artisans like weavers, potters, you had farmers. These jobs too, largely all gone. In the U.S. Today, we should say farmers make up less than 2% of the workforce. Let's move ahead through the changing jobs of the Bronze Age, the Iron Age, the Dark Ages, the Renaissance, and the Age of Exploration. Manoush Zomorodi Age, the same thing happens. Jobs are lost, sure, but more different jobs take their place. Vlad Tenev Too many jobs to count. A lot of them are gone. Any blacksmiths or explorers in this room? I didn't think so. It might come back with space exploration. If you think about it, most of our last names are from jobs that our families no longer do. Potter, Butler, Butcher, Smith. Any Fletchers in the audience? Anyone know what a Fletcher is? I was going over my talk this weekend and my son said, I know what a Fletcher is, Dad. He plays Minecraft. A Fletcher is someone that makes and sells arrows. So if you know someone with that last name, their relatives were arms dealers. Now my point in all of this is job disruption is then an essential quality human evolution. We want work to disappear because it means that we're doing our jobs as humans, making our lives better and easier. So with AI, maybe it's not the job disruption itself that makes us so nervous, but the speed with which it's happening. So why don't we accelerate? We're going to go right through the industrial revolution into the modern era. In the 20th century, a young person in the wake of companies expanding and automating would have found an entirely new menu of jobs that their parents never had access to. So instead of working in a factory, they would have had the selection of a wide assortment of new office jobs. And ([Time 0:39:10](https://share.snipd.com/snip/8108edfd-cfa3-4ffe-ba59-0e8531123675))