Learning how to think just completely changed
Why ChatGPT’s new Study Mode and Google DeepMind’s Genie 3 world builder might reinvent learning, leadership and career development
Automation isn’t cheating—it’s liberation
Every leader fantasizes about having a few extra hours each day to think big thoughts. We long for the quiet time between meetings and emails to connect dots, uncover patterns, and play out scenarios in our heads. Ever since my GE days, I’ve had a whiteboard in my office with a reminder to “stop and think” or “ask bigger questions”. With back-to-back meetings on a range of topics, all requiring a near-instant context switch, even when time is blocked, it’s hard to honor it when competing demands pop up from subordinates, bosses, or colleagues.
Ironically, the technologies people accuse of promoting laziness are delivering exactly the thing we need: time. In a recent conversation at Nationwide, we discussed that as Agentic AI creates time for associates and officers, we have choices in how we should use it. Of course the company will want to soak up some of that productivity goodness by asking people to do more. But, I’m happy to say that we’ve had a lot of conversations about innovation, critical thinking time, team gatherings to explore big ideas or just to connect. All of these offer a new, strategic option to explore that we just simply don’t have the time for today!
Generative AI tools are doing the heavy lifting on coding, research, summarization and administrative busy work. Rather than “cheating,” they allow us to channel our attention toward the strategic thinking and critical judgement that humans do best. The “cheating” hurdle is one we’ve talked about recently as well. Our leadership team has talked about how even officers need to get over the idea that using Generative AI tools is somehow cheating. In the early days of ChatGPT and LLMs, universities and public schools were very worried about student usage of these tools to cheat their way through coursework. Now, there are answers popping up in the industry that will help shift that narrative even at the beginning.
Recent updates from the AI world underscore this shift. OpenAI’s ChatGPT Study Mode encourages learners to engage with material instead of copy-pasting answers. When a user asks about a complex topic, Study Mode asks follow-up questions to gauge the learner’s level and goals, then walks them through the subject step-by-step [1]. OpenAI positions the feature as a way to fight academic misuse; it doesn’t simply spit out solutions, but guides the student toward understanding and allows them to upload exam papers or images for interactive help [1]. This is the opposite of a cheating: it’s a Socratic tutor that prompts deeper reflection.
Meanwhile, Google DeepMind just unveiled Genie 3, a general-purpose world model capable of generating minutes-long, interactive 3D environments from a text prompt [2]. The model remembers what it has generated so simulations stay physically consistent over time [2] and lets users inject “promptable world events” to change the environment on the fly [2]. It isn’t publicly available yet, but the research preview hints at educational uses—Genie 3 could train agents for general-purpose tasks or let learners explore both realistic and imaginary worlds [2] with real-life physics. A separate report notes that the model runs at 720p and 24fps and remembers past interactions, enabling immersive experiences akin to video games while being monitored for ethical use [3]. In other words, AI can now build entire virtual classrooms around us. If you’re a Trek fan, I think we’re seeing the first version of the Holodeck about to be unveiled!
Offloading doing to reclaim thinking
As a data and analytics leader, I’ve watched AI shrink tasks that once ate hours into minutes or even seconds. Drafting reports, generating charts, designing slide decks (well, this one still needs a bit more work…): these can all be automated to a surprising degree. The value isn’t that I “skip the work”; it’s that I can invest my reclaimed time in first-order thinking and strategic leadership. AI becomes the junior analyst who crunches numbers while I consider how those insights feed into broader objectives.
This shift also applies to personal development. Study Mode flips generative AI from an answer engine into a tutor. Instead of handing you an essay, it asks what level of mathematics you’re comfortable with and then scaffolds an explanation [1]. The goal is to encourage engagement, guiding users toward answers rather than serving them up [1]. The feature responds to images too, letting students upload past exam papers so it can help them work through problems step by step [1]. When AI takes over rote tasks and prompts us to think, we build the intellectual muscles that matter most.
From a data perspective, it does make one pause to think about what curation required for enterprise data to drive a professional-world training program. We’ve recently started talking about making sure unstructured data repositories used for training, codes of ethics, company policies, etc. are in fact curated in and of themselves so that we can create a “repository” (which will probably just look like a standard folder on your desktop) that we tune the model to for “official” versions of key policies. We can build a huge ontology and map every document in the organization to it, too, but when you can just create a simplified vault of your latest standard operating procedures, the question becomes, why not take the simple path here? ¯\_(ツ)_/¯
Bringing Study Mode into 3D: Genie 3 and the future of skill-building
Now imagine coupling this Socratic tutor with Genie 3’s immersive world builder. Picture a statistics student stepping into a virtual city where probability concepts are woven into the environment—traffic lights become Bayes’ theorem problems, while the AI tutor probes the student’s reasoning in real time. Or consider a leadership coaching session where you practice difficult conversations with simulated employees. Promptable world events could instantly shift the scenario from calm to crisis, challenging you to adapt on the fly.
Reports highlight that Genie 3 can render whole worlds in real time at 720p and 24fps for minutes at a stretch, remembering interactions so objects stay where you left them and letting users change weather or summon new elements with simple text commands [3]. This emergent visual memory means that learners could revisit the same virtual space and find it consistent; their actions would matter. Combine that with Study Mode’s Socratic questioning and you have a fully interactive, adaptive classroom that responds to both your answers and your actions.
Such environments could reinvent professional development. Today’s leadership training often relies on role-playing and case studies. Tomorrow’s could immerse managers in dynamic, AI-generated worlds where they practice decision-making, systems thinking and empathy. As DeepMind researchers note (via reporting), Genie 3 is designed to train agents for general-purpose tasks and pushes them to learn from their experience [2]. Why not treat ourselves as agents in our own learning?
Final thoughts: embrace the paradox
AI is often criticized for making us lazy or for hollowing out human expertise. But if we use it thoughtfully, it can do the opposite. Automating administrative drudgery frees us to focus on judgement, creativity and relationships. Socratic AI tutors push us to think rather than regurgitate. Immersive world models promise experiences that no textbook could match. And when we bring these tools into our leadership practice, we may finally get what we’ve always wished for: more time to think deeply about what matters.
The paradox is that the more we let AI do for us, the more we’re forced to confront ourselves. That’s not cheating—that’s growth.