A (Somewhat) Meta Analysis of the impact of DeepReserch AI tools... and a bit about what we should do about it.
Research, thought work, and specialized knowledge are now accessible through drag-and-drop tools. To better understand these tools, I adjusted my usual research methodologies for the blog you’re currently reading.For background, I love research—it’s a huge part of my writing process. I typically examine at least 20-30 pieces to find usable references. To get there, I have to sift through many more. I enjoy this process because it lets me explore different facets of a topic. The downside? It’s labor-intensive. There’s a lot of culling and evaluating sources. Ironically, I almost never use other blogs, opting instead for academic and news-based resources. For this piece, I fed the following prompt to the free trial version of Google Gemini’s Deep Research tool:I want to understand the implications of tools like deep research on society. With a particular focus on the sociopolitical ramifications of such tools within a society that offers premium wages for thought work. What scientific and academic studies have explored the potential impact of these tools on work and work culture? I also want to understand what impact these tools will have on labor costs overall. please include information about how companies are currently adopting these tools and what impact it is having on the current workforce.
Candidly, this is a lazy prompt. It's overly vague, filled with jargon, and lacks constraints. I wanted to simulate how a non-AI practitioner might query the product. The results were fairly impressive.
From a UX perspective, I appreciated that the reasoning parameters were transparent. Gemini explained, point by point, how it interpreted the prompt and applied that interpretation to its research. It also provided a full list of sources used to explore each facet of the topic.
It examined far more blogs than I typically would, layering those with academic sources and sites demonstrating industry expertise—like SHRM, an HR platform it used to assess labor impacts. If you're interested, the tool produced a full research paper, shared below.
Honestly, it's a good read—accessible and informative. I’ve highlighted the elements that most interest me and significantly inform this piece. I’ve also added commentary to key sections to show how I interpreted the curated research and synthesized it into something new.
But this text—the one you're reading now—isn't about the topic I prompted. It's about how we respond to the creation of these tools and how we integrate them into our lives. Hence, the meta approach. The goal here is to illustrate how to incorporate AI-generated outputs into novel, human-directed work, just as I would with “manually derived” research.An Evolve or Perish moment
Learning to use these tools goes beyond rote skills like "how to prompt" or "how to train." It’s fundamentally about how to critically interpret outputs and build upon them. While tools may democratize access to thought work or Ph.D.-level research, the operator still needs to discern, engage, and critique. Human capabilities like discernment, critique, and intellectual synthesis are irreplaceable and remain largely undemocratized.
If your understanding of a topic is surface-level, these tools might help you fake expertise for a while. But in the long term, you’re proper fucked. When research is handed to you, superficial insight no longer cuts it. You’re now expected to extend beyond what’s readily available and generate real value from your perspective.
Crucially, you must cultivate the ability to critique and bring a critical valence to your prompting. While sources shape the perspectives AI presents, there’s often an inherent positive bias. Reckoning with ideas curated by tools like these isn't just about asking the right questions. It's about the human intellectual task of dissecting and synthesizing. In work contexts, critical thinking remains the most potent form of human magic, not only to generate usable outputs but to build something new upon them.
Fortifying Our Economy in the Face of Technological Disruption
Usman Sheikh, a must-follow on LinkedIn, recently posted a valuable take on this topic in response to OpenAI’s unveiling of its $20,000/month agentic research product. I’ve included it here, so there’s no need to rehash his excellent analysis.
Make it stand out
That said, a bit of context:The U.S. economy has been unpredictable over the past three months, recalibrating across several fronts. We’ve seen a frozen job market, widespread uncertainty, and high inflation. All while both public and private sectors push for layoffs and invest deeply in digital labor.
While I’m not apolitical, my focus here is the impact of technology on political and social norms. This space strives to offer an apolitical analysis of potential policy responses to the conditions outlined above.
As a society, we must make hard choices about how we want the economy to function in this new era. As practitioners and experts, we need to help non-SMEs imagine new economic paradigms as old norms collapse. It’s socially irresponsible to leave these decisions in the hands of a few powerful, self-interested parties.
We’ve faced such moments before—and we can learn from how we responded.
What We Can Learn from FDR
If you remember high school history, a few key facts about the New Deal likely stand out:It marked the end of laissez-faire economic policies and embraced Keynesian principles, focusing on government spending and fiscal stimulus.
It birthed the modern labor movement. The Wagner Act guaranteed workers the right to organize, shifting power dynamics between employers and employees.
It didn’t end the Great Depression (WWII did), but it forged the modern economic era.
It was a moment of reckoning when we recognized that the old world was dying and began building the foundation of a new one. We find ourselves in such a moment again.Again, we are in need of strong regulation and improved labor protections. There’s also a compelling case for strengthening the social safety net and exploring solutions like universal basic income, but these are ideas that are far from being de rigueur.
Instead, we’re making deep cuts where we might need reinforcements. Again, this is not a political argument but a call to reconsider whom our economy is designed to serve as we design ever more powerful AI systems. No one is asking for a welfare state, just a focused recalibration of existing structures in light of what’s coming...well, what’s here.Defining Cognitive Protections
Unsafe at Any Speed gave us seatbelts. Silent Spring gave us the EPA. The Jungle gave us food safety laws. Each revealed dangers society had previously ignored.
Since the rise of social media in the early 2000s, we've struggled to enact meaningful policy around algorithmic systems. We are, metaphorically, pantsless in the face of AI's commercialization.
Despite mounting evidence that prolonged use of AI tools can degrade cognitive abilities, we still lack consensus on how to constrain or refine their use. The market’s brute-force approach has led to widespread labor cuts and unfocused applications that rarely deliver on AI’s promise of human augmentation.
We’re failing to define these systems, and we’re not establishing effective norms for human–AI teaming. The problem is systemic, compounded by market pressure to “do something” without clearly defined paths to value.
The rise of deep research tools threatens the very institutions that might help solve this problem, institutions like consultancies and academia. Yet these same institutions remain our best existing mechanism for guiding organizations and individuals toward equitable, value-driven solutions. If they’re to survive, they must evolve beyond surface-level guidance.
Deep research tools give consultancies a chance to streamline and operate with leaner teams. But there’s a brief window during which they must elevate those within their ranks who are capable of systems-level thinking and let go of middling talent. Business as usual will lead to inevitable decline.
AI is no longer "on the horizon." We're not “experimenting” anymore. We are evolving with it. Every sector of society must adapt—or perish.
Key Takeaways
The era of superficial thinking is over. Easy-to-use research tools have shifted expectations of what qualifies as thought work. Critical thinking is now a prerequisite. Deep knowledge and the ability to synthesize information into original perspectives are baseline skills in today’s labor market.
We can look to the past for inspiration on how to fortify our economies in the face of a new order. Society cannot ignore this task, as the disruption to come will be significant and wide-ranging.
Consultancies, academic institutions, and policymakers must swiftly restructure to guide more of society through this transformation. If they don’t, they’ll face extinction.
Disclaimer: The opinions expressed in this blog are my own and do not necessarily reflect the views or policies of my employer or any company I have ever been associated with. I am writing this in my personal capacity and not as a representative of any company.
This article was edited with the help of Editorial AI .
AI Has a Use Case Problem—Because It Also Has a Practitioner Problem
(And on a Macro Level, Society Has an Imagination Problem)
Let’s dig in.This might be surprising coming from an AI practitioner, but I am a bit of a geek. And I mean that in the true Patton Oswaldian Otaku rant sense of the word. I love my Stars—both Wars and Trek—and probably any other nerdy piece of IP you can think of. So, my apologies to the uninitiated, but we’re going to get a little geeky this week. I promise if you stick with me, it’ll all make sense in the end.The Link Between Imagination and Creation
What we can imagine directly impacts what we choose to create. Star Trek is a perfect example. There’s no shortage of think pieces outlining how the show influenced modern technology—whether it’s this article, this video, or this full list. You get the point.And it’s not just about physical objects. Fiction also shapes broader social concepts, like the normalization of certain marginalized groups.From the iPad to Bluetooth headsets, sometimes a key step in technological advancement is seeing proto-versions of it in art. This is probably why society is so obsessed with humanoid robots. Robots Are Boring
Okay, hear me out. Of course, I have my favorite fictional bots—R2D2 sits at the top of my list because I have two eyes and a heart. But in this current moment, I can’t help but offer up an eyeroll and a deep sigh every time I see yet another demo of the “latest humanoid robot.”Most—if not all—fall squarely into uncanny valley territory. Take this gem from Clone Alpha, or any of these female-presenting bots that, unsurprisingly, almost universally display personalities programmed by men.And then there’s the hype cycle around Figure 1, the humanoid robot from OpenAI. Ultimately, what is cool about Figure 1 is not its humanoid appearance or the servile nature of its tasks but the individual pieces of tech that make it up. Speech-to-text reasoning, persistence, and object recognition are cool. However, using all of those things to create a race of servants is not cool at all. Pardon me for saying it out loud, but humanity can do better.It disturbs me that so much time, money, and effort is spent forming the most advanced technology we’ve ever created into what is essentially a new underclass. Sure, the idea of a domestic bot—something that does your laundry, dishes, and cleaning—is appealing. There’s even an argument that automating household chores could positively impact gender relations, given that women still do the majority of domestic labor.But do we really need a humanoid servant to accomplish that? Wouldn’t it make more sense to use smart objects, such as Roombas, smart fridges, and other integrated IoT devices?What is the key difference between these approaches? Besides a couple billion dollars in R&D, the fundamental difference is this:An IoT approach creates a constellation of human-operated tools that facilitate social and behavioral change.A humanoid approach replaces human labor with…a different labor force.
Not to mention that humans should clean their own spaces. It’s good for mental health, strengthens our connection to our environment and families, and even has a positive impact on mood. Human beings take a long time to evolve, and the creation of a novel piece of technology doesn’t mean we somehow change. Things that help us manage our nervous system reactions to our spaces are important to preserve. The humanization of AI feels inevitable because we have told ourselves that it is inevitable for decades now. The first use of the term robot was in a 1921 Czech play R.U.R. (Rossum's Universal Robots). In the play, these robots are used to replace human work. They eventually rise up and wipe out the human race. This is a common theme in robot/AI-related art from Asimov to The Matrix. The concept of a robot has long been used as a cautionary tale, critiquing the human desire for domination and subjugation.We’re so hypnotized by past portrayals of AI that we’re not fully exploring what AI could actually do. Humanoid robots are overrepresented in our cultural imagination, and as a result, they dominate our real-world AI ambitions—at the expense of more relevant and socially helpful applications.AI in the Workplace: Misguided Investment and Poor Execution
Credit: Tom Fishburne
Deloitte reports that generative AI, specifically, attracts the most investment across different sectors in IT, operations, and customer service use cases. These investments make sense for the current class of AI tools, which are typically aimed at logistics, code writing, and data handling. Yet 80% of AI projects fail because they are not rooted in workstreams that create value. Consultants overwhelmingly train their clients to chase what is new and next instead of considering specific business needs. Worse, even when good use cases are pursued, stakeholders often misunderstand or misinterpret them. Most importantly, most organizations don’t have sufficient frameworks, structures, or protocols to accommodate the use of AI tools. This is a disaster across the board. This results in bad investments, frustrated stakeholders, and a whole lot of wasted potential.What can companies, consultants, and project stakeholders do to help mitigate these issues? Stay educated about AI. The field moves fast, and workers need to use these tools regularly to understand their limitations and refine their outputs.Root use cases in value. Instead of chasing trends dictated by CEOs or consultants, companies should focus on what supports their workflow.Evolve ways of working. To ensure sustainable integration, organizations need clear review structures, governance policies, and AI management roles.
We need more Toys
Most technology starts out as a toy—because humans learn best through play.Unlike robots, our nervous systems are a key aspect of cognition. Passive states of engagement positively impact education and adaptation to new technology.But AI has yet to have its “cool toy” moment. So far, AI commercialization has been overwhelmingly work-focused. Beyond entertainment, we need imaginative, socially beneficial applications of AI. Here are a few ideas: A Language Ark
Joel Satore’s Photo Ark is a breathtaking and inspiring masterpiece that subtly draws attention to the role of climate change on ecosystems. The National Geographic photographer’s efforts aim to capture images of every living animal; they have currently photographed 16,000 creatures to date. If you have never seen it, please check it out.LLMs create the potential for similar efforts to preserve dead, dying, or rarely encountered languages. Dying languages have fewer than 10 living speakers. Rare languages have more speakers but are not often encountered because they are spoken by people in remote areas or from isolated social groups. Queens, New York, for instance, not only has the most languages spoken in the world but also boasts the highest concentration of dying languages. The New York Times recently published an outstanding article on this subject, which I recommend everyone read. Languages are not just words; they carry knowledge of concepts that sometimes exist only within the culture of their origin. This application showcases AI’s greatest strength: its ability to parse and vectorize language, which could be both inspiring and transformative for the world. Nature Glasses:
We already have smart glasses that can help us contextualize information in a heads-up display. However, the use cases associated with these tools are work-related or tied to commerce in urban environments. Why not bundle AI insight into these tools to help us better understand the natural world? Tools like Picture This identify flowers, Merlin uses smartphone features to help identify birds, and Night Sky helps identify constellations. Why not create smart glasses that help us explore the natural world? Rescue Logistics:
With the LA fires still top of mind, it makes sense to consider how AI would help with the logistics associated with responding to natural disasters. Owing to climate change, we are likely to need such tools on an ongoing basis. Imagine being able to accurately assess where food resources are and manage the logistics of distributing them to needy populations. Or coordinating overall mitigation efforts with machine precision?
AI technologies are powerful and useful. When combined with the power of the human imagination, they can make wondrous things possible. Key Takeaways
We need to imagine more from technology to fully realize its potential. Pop culture, art, and our collective human creativity directly influence what we choose to build and believe in. Right now, AI representations are narrow and stagnant—dominated by humanoid robots and outdated sci-fi tropes. We need bigger, bolder ideas to inspire real innovation.Organizations looking to drive productivity and growth through AI need to do a much better job of best-fitting use cases. AI investment needs massive reform, and consultants need to stop pushing sales-driven hype instead of real education. Fixing this will unlock actual value instead of just more failed projects.Society needs inspiring AI applications that drive real engagement and exploration. AI’s adoption—and its ultimate impact—won’t be shaped by corporate boardrooms alone. We need more play, more wonder, and more creativity in how we interact with this technology.
Some of the most powerful and transformative technologies in history didn’t start out as corporate solutions—they started as toys, art, and experiments in human curiosity. AI needs that moment. If we let imagination lead the way, the results could be extraordinary.Disclaimer: The opinions expressed in this blog are my own and do not necessarily reflect the views or policies of my employer or any company I have ever been associated with. I am writing this in my personal capacity and not as a representative of any company.
About this Article
As a graduate of the University of Missouri School of Journalism, I understand the value of strong editorial oversight. While I crafted the initial draft of this article, I recognize that refining complex narratives benefits from a meticulous editing process.
To enhance clarity, cohesion, and overall readability, I collaborated with The Editorial Eye, a ChatGPT-based AI designed to function as a newspaper editor. According to the tool, its refinements aimed to “enhance readability, strengthen argument flow, and polish phrasing while preserving the original intent.”
However, the editing did not stop there. After reviewing the AI-assisted revisions, I conducted a final pass to ensure the article accurately reflected my voice and intent. The AI did not generate new ideas or content; rather, it helped refine my original work.
What you see here is the product of a thoughtful collaboration between human insight and AI-driven editorial support.