Don’t Blame the Black Box: Why We Avoid AI Explanations

Posted on January 6, 2026 by Eugene Ivanov

There’s a Russian proverb that cuts straight to the heart of human nature: Having an ugly face, don’t blame the mirror (На зеркало неча пенять, коли рожа крива).

We like to blame LLM models for their lack of transparency. Calls for Explainable AI get louder every single day. Politicians demand it, regulators require it, and businesses claim it’s essential for winning consumer trust. And yet, when an AI tool provides us with a transparent explanation of its reasoning, we tend to ignore this explanation, especially if it makes us uncomfortable or, worse, if it threatens to undermine our financial gains.

The Loan Officer Experiment: Seeking Predictions, Avoiding Truth

New research from Harvard Business School highlights this unsettling truth in two experiments.

The first experiment placed participants in the role of loan officers at a private U.S. lender. Their task was straightforward: allocating a real, interest-free $10,000 loan between two unemployed borrowers. An AI system had classified one borrower as low risk and the other as high risk. Participants could access the AI’s risk predictions and also choose whether to view an explanation of how the model reached its assessment.

The results were striking. Roughly 80% of participants eagerly accepted the risk scores: they wanted the AI’s predictions to help them make profitable decisions. But only about 45% chose to see the explanations. The gap widened dramatically when participants’ financial incentives were aligned with the lender’s interests: they earned more money if the loans were repaid. (Lender-aligned participants were about 10 percentage points more likely to skip explanations than neutrally compensated participants.) These participants were even more likely to seek out predictions, but significantly more likely to avoid explanations, particularly when told that those explanations might involve considerations of race and gender.

The pattern was clear: when financial incentives conflict with fairness concerns, people don’t just make questionable decisions; they strategically avoid information that would force them to confront the ethical dimensions of those choices.

Critically, this wasn’t about disliking extra information in general. When race and gender information was removed from explanations and replaced with arbitrary technical details, the gap in explanation avoidance between different incentive conditions almost vanished. People weren’t shunning explanations as such; they were avoiding what the explanations might reveal about discrimination and their own profit-maximizing behavior.

The Risk Experiment: Failing to See What Helps

The second experiment removed moral conflicts entirely to focus on pure decision quality. Here, participants evaluated a loan application that an AI had labeled “high risk” because of a two-year employment gap in the borrower’s work history. The researchers first asked participants how much they would be willing to pay for an explanation that would reveal whether the employment gap was indeed the primary driver of the AI’s high-risk classification.

Then came the crucial twist. Participants received free private information: the employment gap resulted from the borrower pursuing a full-time professional certificate, a benign reason that shouldn’t increase credit risk (unlike, say, a job termination). This private information should have made the AI explanation significantly more valuable: If participants knew that the AI’s high-risk label stemmed from the employment gap, and they also knew the gap resulted from pursuing education rather than being fired, they could integrate both pieces of information to reach a more accurate risk assessment.

Logic suggests that after receiving this private information, participants should value the AI explanation more highly, not less. But that’s not what happened. When asked for the second time about their willingness to pay (that is, after receiving the private information about the certificate) valuations actually dropped by 26%. People systematically failed to recognize that the explanation would help them integrate their own knowledge with the AI’s output to make a better decision.

Only when researchers explicitly walked participants through the logic, spelling out exactly how the private information and AI explanation could be combined, did the valuations increase. This revealed a novel behavioral bias: people don’t naturally see when explanations would be most useful to them, even when there’s no moral conflict involved.

The Black Box We Refuse to Open

We often complain that LLM models are like a black box and criticize AI Labs for creating them. The metaphor has become ubiquitous in debates about artificial intelligence: mysterious algorithms making consequential decisions while we’re left to wonder what’s happening inside.

But this research reveals an uncomfortable irony. When an AI algorithm gives us an opportunity to open the lid of that black box—to peer inside and understand its reasoning—we hesitate. We look away. Sometimes it’s because we’re lazy and don’t want to make the cognitive effort. More often, it’s because we don’t want to see what’s there.

In high-stakes decisions spanning credit, hiring, pricing, healthcare, and safety, people may eagerly consume AI predictions while quietly avoiding the explanations that would expose uncomfortable trade-offs or discriminatory patterns. That avoidance can skew outcomes, undermine fairness, and create hidden risks for every organization. Meanwhile, even well-intentioned professionals may systematically under-invest in explanations that would improve their forecasting by helping them combine their domain expertise with AI insights.

Building transparent AI systems is necessary but not sufficient. The real challenge isn’t engineering better explanations or making black boxes more transparent; it’s ensuring that people actually use the transparency that’s already available. Organizations must design decision-making environments and incentive structures that encourage opening the lid, even when what’s inside might be uncomfortable.

Because the black box isn’t the problem. The problem is our unwillingness to look inside.

Posted in AI | Tagged , , , , | Leave a comment

The AI Paradox: Why SMEs Might Be Losing Ground Just When They Thought They’d Caught Up

Posted on December 3, 2025 by Eugene Ivanov

Three months ago, I declared AI the great equalizer for small and medium enterprises. Today, I’m not so sure. In fact, I’m worried we might be celebrating prematurely — and that the very technology promising to level the playing field could actually widen the gap between SMEs and their larger competitors.

The September Dream: AI as the Great Democratizer

Back in September, I published “Unlocking SME Innovation: Why AI-Based Problem-Solving is the Great Equalizer,” where I celebrated what seemed like a generational opportunity for SMEs. My argument was straightforward and optimistic: advances in AI, particularly large language models, were dramatically reducing the cost of domain expertise. Tools like ChatGPT, Claude, and Gemini were giving everyone access to knowledge that just years ago was exclusive to large, resource-rich organizations.

The promise was intoxicating. SMEs could now engage in sophisticated scenario planning, competitive analysis, and innovation forecasting — capabilities previously reserved for corporations with dedicated strategy teams. Where SMEs once waited weeks for external consultants or struggled in isolation, AI provided immediate domain expertise, alternative approaches, and consequence analysis before committing resources.

I believed — and still want to believe — that agility can become more valuable than resources, and creative problem-solving can trump bureaucratic processes. But recently, I’ve encountered a paradox that fundamentally challenges this optimistic vision.

The Knowledge Dichotomy: When More Becomes Less

To understand the paradox, we need to distinguish between two types of knowledge: explicit and tacit.

Explicit knowledge is information that can be easily codified, documented, and transferred. It’s the data in your reports, the insights in your dashboards, the processes in your manuals, and the analyses in your presentations. This is precisely what AI excels at generating. LLMs can analyze market trends, produce competitive intelligence, create strategic frameworks, and synthesize information from vast datasets — all at unprecedented speed and minimal cost.

Tacit knowledge, by contrast, is deeply embedded in experience, intuition, and context. It’s the expert judgment honed over years of practice, the creative problem-solving that comes from pattern recognition across multiple situations, the ability to read a room and build relationships, and the organizational culture that shapes how decisions get made.

Here’s where AI turns the tables: by making explicit knowledge abundant, cheap, and universally accessible, AI simultaneously commoditizes it. If your AI tools can generate sophisticated market analysis, so can your competitors.’ If you can produce detailed competitive intelligence reports, so can everyone else in your industry. Explicit knowledge, once a source of competitive advantage, becomes a common baseline rather than a differentiator.

And as explicit knowledge loses its strategic value, tacit knowledge becomes the critical differentiator. The paradox is complete: AI democratizes explicit knowledge while elevating the importance of the very thing that can’t be democratized — human experience, judgment, and intuition embedded within organizations.

The SME Threat: Winning the Battle, Losing the War

This paradox poses a particularly acute threat to SMEs, and it’s one I didn’t fully appreciate in September.

Yes, SMEs can now generate the same volume of explicit knowledge as their larger competitors. They can produce equally sophisticated analyses, reports, and strategic frameworks. “We have access to the same knowledge as you guys!” they might justifiably claim.

But here’s the problem: explicit knowledge is only half the equation—and increasingly, it’s the least important half.

Large organizations possess something SMEs often lack: a critical mass of accumulated tacit knowledge. They have teams of experienced professionals who’ve navigated multiple market cycles, managed countless customer relationships, and learned through trial and error what works and what doesn’t. They have established decision-making processes refined over decades, institutional memory that prevents repeated mistakes, networks of expertise that span functions and geographies, and organizational cultures that know how to translate insights into execution.

This tacit knowledge infrastructure is what turns data into decisions, and decisions into results. It’s the interpretive layer that determines which AI-generated insights matter and which don’t, the judgment that knows when to act boldly and when to proceed cautiously, and the execution capability that transforms analysis into competitive action.

So, here’s the cruel irony: by democratizing explicit knowledge, AI may widen the gap between SMEs and larger players. SMEs gain access to knowledge abundance but lack the tacit knowledge infrastructure to leverage it effectively. They’re drowning in insights but starving for wisdom.

Fighting Back: Building Tacit Knowledge at Scale

Should SMEs surrender to this paradox? Absolutely not. But they need to be strategic about how they compete in an AI-augmented world.

First, SMEs must recognize that their competitive advantage won’t come from AI-generated knowledge itself — it will come from how they apply that knowledge through their unique tacit knowledge capabilities. This requires intentional investment in building organizational wisdom, not just accessing information.

Second, SMEs should focus on what they can do better than large organizations: developing deep, contextual understanding of their specific customers and markets. Large companies have breadth; SMEs can have depth. Know your customers not just through data, but through relationships, repeated interactions, and intuitive understanding of their unstated needs.

Third, create tight-knit, high-trust teams where tacit knowledge flows naturally. In smaller organizations, this is easier to achieve than in large bureaucracies. Use this structural advantage to build learning cultures where experience is shared, mistakes are discussed openly, and collective judgment improves continuously.

Fourth, implement deliberate knowledge transfer mechanisms — mentoring programs, case study discussions, post-project reviews — that capture and disseminate tacit knowledge across your organization. Don’t let experience remain siloed in individual heads.

Finally, use AI strategically to augment your tacit knowledge, not replace it. Let AI handle the explicit knowledge generation: data analysis, report creation, and pattern identification. This frees your people to focus on interpretation, judgment, and creative application — the tacit knowledge work where you can still differentiate.

The Window Is Closing

The adoption window for AI is compressing rapidly. Following historical patterns, we likely have only 3-4 years until peak adoption in 2028-2029. SMEs that spend these precious years simply celebrating access to AI-generated explicit knowledge will find themselves competitively disadvantaged despite being technologically enabled.

The winners will be those who recognize the paradox and act on it now: embracing AI for what it does best while urgently building the tacit knowledge capabilities that AI cannot replicate.

The great equalizer might not be so equal after all. But for SMEs willing to play a different game — one focused on wisdom rather than just information — the opportunity remains extraordinary.

I’m grateful to Daniel Martinez Villegas, whose recent presentation at the Berkeley Open Innovation Seminar drew my attention to the explicit vs. tacit knowledge dichotomy.

Posted in AI, Innovation | Tagged , , , , , , , , , , , | Leave a comment

Putting the Cart Before the Horse: What We’re Getting Wrong About AI

Posted on November 10, 2025 by Eugene Ivanov

The debate about artificial intelligence has become exhaustingly predictable.

On one side, we have doomsayers who celebrate every misstep—a misdrawn map of Europe, a miscounted number of r’s in “blueberry”—as proof that AI is fundamentally flawed. The word “hallucination” has been weaponized to dismiss technology that, despite its imperfections, has made extraordinary strides in reliability. On the other side, we have enthusiasts, armed with an ever-expanding toolkit of specialized models and applications, who rush to integrate AI into every conceivable business process.

Both camps, I would argue, are missing the point.

The skeptics’ position barely warrants discussion. Yes, AI makes mistakes. So do humans—with alarming regularity. The relevant question isn’t whether AI is perfect, but whether it’s useful. And on that measure, the evidence is overwhelming. Major language models have dramatically reduced their error rates, and their capabilities continue to expand at a pace that would have seemed impossible just years ago. Dismissing this technology because it occasionally stumbles is like rejecting automobiles because they can’t navigate every dirt path that a horse can.

But here’s where it gets interesting: even among those who embrace AI’s potential, most are approaching its implementation backwards. I call this the technology-centric trap. The thinking goes something like this: “We have these amazing AI tools available. Which of our existing business processes can we automate with them?” It’s a natural question, especially given the dizzying array of AI applications flooding the market, each promising to revolutionize some aspect of operations.

The problem is that this approach assumes our current business processes are fundamentally sound—that they just need a technological upgrade to run faster and cheaper. But what if the processes themselves are the problem? What if they’re outdated, inefficient, or built on assumptions that no longer hold in today’s environment? Bolting AI onto broken workflows doesn’t fix them; it just automates dysfunction at machine speed.

The correct sequence is elegantly simple, though harder to execute: identify the problem first, then find the solution. Not the other way around.

This isn’t theoretical musing. My experience with crowdsourcing taught me this lesson clearly. Successful crowdsourcing doesn’t start with assembling a crowd and asking what they can solve. It starts with identifying a specific problem, tracing it to its root cause, and defining it with precision. Only then do you present it to potential solvers. Skip those preliminary steps, and you’ll get solutions to the wrong problems—or no workable solutions at all.

The same principle applies to AI integration. Before asking which AI tools you should deploy, ask: What processes are genuinely holding us back? Where are the bottlenecks that constrain growth? Which workflows were designed for a different era and have simply persisted out of habit? These questions require honest, sometimes uncomfortable introspection about how your organization operates versus how it should operate.

Only after answering these questions does it make sense to survey the AI landscape. If appropriate tools exist, deploy them. If they don’t, consider building them or adapting what’s available. But the technology choice flows from the problem definition, not the reverse.

IBM’s recent paper on AI agent architecture makes this point compellingly. Their analysis reveals that many AI agent deployments stall after the pilot phase not because the technology fails, but because organizations are trying to force-fit advanced AI onto fundamentally broken workflows. Technology works fine; the underlying processes don’t.

This isn’t about being anti-technology or advocating for needless delays. It’s about being strategic. AI offers unprecedented opportunities to reimagine how work gets done, but only if we’re willing to question the status quo first. The businesses that will truly benefit from AI aren’t those that deploy the most tools the fastest. They’re the ones that take the harder path: examining their operations critically, identifying what needs to change, and then—and only then—leveraging AI to build something better.

The future belongs not to those who automate the present, but to those who redesign it first.

Posted in AI, Crowdsourcing | Tagged , , , , , , , , | Leave a comment

Unlocking SME Innovation: Why AI-Based Problem-Solving is the Great Equalizer

Posted on September 7, 2025 by Eugene Ivanov

In mid-October 2009, I was visiting with a client, a large, Midwest-based paint and coating manufacturing company.

As part of their product development process, the company’s engineers built a powerful outdoor pump to paint industrial buildings. The pump worked beautifully in indoor testing, but when the engineers tried to use it outdoors, the pump started to clog frequently, making it essentially useless. The client wanted me to help them run a crowdsourcing campaign aimed at redesigning the pump.

When speaking with my counterparts at the client’s innovation group, I pointed out to them that the indoor and outdoor conditions they used to test the pump weren’t identical: the indoor testing was done during the summer, with the temperature even in the air-conditioned lab often reaching the mid-seventies, while the outdoor temperature in the Midwest at this time of year rarely hit the 60°F mark. Could it be that the clogging was somehow caused by the temperature shift?

My hunch turned out to be correct. The problem was not the pump design. The problem was the paint: it was rapidly becoming viscous with a small drop in temperature, causing the pump to clog. The engineers fixed the problem by simply adjusting the paint formulation.

As an innovation manager, I like to remind my clients that the most important part of the problem-solving process is to correctly define the very problem they’re trying to solve.

The sad reality is that many large organizations, both corporate and non-profit, fail to identify the root cause of their problems. Instead, they immediately start looking for something—anything!—that may look like a solution.

To me, this is equivalent to taking Tylenol to relieve a headache even before knowing what caused it: hangover, mild cold, chronic migraine, or advanced glioblastoma.

The situation is even worse for small- and mid-sized companies (SMEs). They’re under constant pressure to innovate, but often lack the dedicated innovation departments, large budgets, and internal resources that their larger competitors rely on. While traditional consulting firms primarily cater to enterprise-level clients, SMEs are often left underserved, leaving their internal problem-solving capabilities ad hoc at best.

The AI Revolution: A Generational Moment for SMEs

Advances in AI, particularly large language models (LLMs), dramatically reduce the cost of domain expertise. By using tools like ChatGPT, Claude, or Gemini, everyone can now tap into knowledge that just a few years ago was accessible only to large and resource-rich organizations.

This presents a generational opportunity to level the playing field. The AI-based tools can stimulate the creative process, energize problem-solving, and support decision-making at SMEs with unprecedented speed and affordability.

What we’re witnessing isn’t simply an upgrade to existing business tools—it’s a fundamental shift in how problems get solved. Consider the cognitive cleanup that AI enables: where SMEs once struggled to sift through mountains of data, identify patterns, or generate multiple solution pathways, AI tools can now process complexity in real-time, offering structured thinking frameworks and systematic approaches to innovation challenges.

This transformation enables real-time business unblocking. When an SME faces a technical hurdle, market challenge, or operational bottleneck, AI tools can immediately provide relevant domain expertise, suggest alternative approaches, and help teams think through consequences before committing resources. The days of waiting weeks for external consultants or struggling in isolation are rapidly ending.

The emergence of problem-solving intelligence through AI represents more than efficiency gains—it’s about democratizing strategic thinking itself. SMEs can now engage in sophisticated scenario planning, competitive analysis, and innovation forecasting that were previously the exclusive domain of large corporations with dedicated strategy teams.

What makes this moment truly generational is the compound effect: as AI tools become more sophisticated and SMEs become more adept at leveraging them, the competitive advantages traditionally held by larger organizations begin to erode. Agility becomes more valuable than resources. Creative problem-solving trumps bureaucratic processes.

The Future of Innovation Services

This is also the moment to redefine traditional consulting by combining human expertise with AI tools and bringing cutting-edge innovation practices to SMEs across industries. It’s time to introduce AI-augmented innovation services for SMEs.

The new paradigm isn’t about replacing human insight with artificial intelligence—it’s about amplifying human creativity and judgment with AI’s processing power and knowledge synthesis capabilities. This hybrid approach enables SMEs to punch above their weight class, competing not just on price or niche expertise, but on the quality and speed of their innovation processes.

It’s not a transient trend. It’s a blueprint for the next generation of SME decision-making. The organizations that embrace this shift now will find themselves equipped with sustainable competitive advantages that compound over time, while those that hesitate risk being left behind in an increasingly AI-augmented business landscape.

Posted in AI, Innovation | Tagged , , , , , , , , , | Leave a comment

The Brainstorming Renaissance: How GenAI Tools Are Rewriting the Rules of Creativity

Posted on May 24, 2025 by Eugene Ivanov

What if the best idea in your next big innovation meeting didn’t come from your star designer, but from a chatbot?

This isn’t a futuristic thought experiment; it’s happening now. Generative AI tools like ChatGPT, Midjourney, and Stable Diffusion are infiltrating brainstorming sessions, product design sprints, and even poetry readings. They’re not just helping — they’re outperforming human contributors on key metrics like speed, idea quality, and production cost.

As the ideation landscape is redrawn, it raises profound questions: Are AI-generated ideas better than human ones? Who benefits most from these tools — the seasoned expert or the curious novice? And more provocatively, is this the death of creativity, or its long-overdue rebirth?

Let’s unpack this creative renaissance in two acts.

Act I. GenAI vs. Human Brains: The Battle of Ideas

Quality, Novelty, and Feasibility: The Metrics That Matter

The old belief that “creativity is uniquely human” is quickly eroding. A landmark 2023 study by Girotra and colleagues compared ideas generated by ChatGPT-4 with those brainstormed by students at an elite university. 

The task? Inventing commercially viable products. The results? Staggering.

ChatGPT-4 produced ideas with higher average quality, measured by consumer purchase intent. It also dominated the high-performance tier — 35 of the top 40 ideas came from the model, not the humans. And it did all this at 40 times lower cost than its human counterparts.

Similarly, Meincke et al. (2024) showed that when GPT-4 was fed a few high-quality examples (a technique known as few-shot prompting), its outputs significantly outpaced those from human ideators across multiple dimensions of perceived value, though humans still edged out the machine on idea novelty.

This novelty gap has consistently surfaced across domains. In innovation tasks, artistic expression, and even scientific ideation, humans tend to produce slightly more novel ideas. But here’s the twist: being novel doesn’t always mean being better.

In real-world innovation, novelty without feasibility might be just noise. That’s where GenAI shines — balancing utility with surprise. In the words of Joosten et al. (2024), AI-generated ideas often have higher customer benefit and overall value, even when they are only moderately novel.

Similar things happen in the art world. When human evaluators were asked to judge whether a poem was written by a human or ChatGPT-3.5, they failed to tell the difference, and often preferred the AI version. The reason? AI poetry was rated higher on rhythm and beauty, two key markers of aesthetic impact.

The creative playing field isn’t just leveling — it’s shifting.

Speed and Cost: The Unfair Advantage of GenAI

Creativity has always come at a cost: time, energy, expertise. Generative AI blows this equation wide open.

In a 2024 study by Boussioux et al., AI generated high-quality business ideas at a fraction of the time and cost compared to human crowdsourcing. Human-generated solutions cost $2,555 and 2,520 hours. GPT-4 produced comparable (and in many cases better) ideas in 5.5 hours and for only $27.

In artistic domains, the same pattern holds. Zhou and Lee (2024) analyzed over 4 million artworks and found that artists using GenAI tools experienced a 25% increase in productivity and a 50% boost in engagement metrics like likes and shares. GenAI didn’t just amplify quantity; it elevated quality, especially when human artists actively filtered and curated the outputs.

But this productivity surge comes with a subtle risk: homogenization. Studies consistently show that GenAI outputs, particularly when used en masse, tend to be more similar to each other. The diversity of ideas — that raw, unpredictable chaos of human thought — gets smoothed out by the statistical instincts of the machine.

Prompt engineering can mitigate this to an extent. Techniques like chain-of-thought reasoning or persona-driven prompts have shown promise in boosting AI’s creative variance. But for now, GenAI is a volume weapon, not a chaos engine.

Act II. Who Gains More? Novices vs. Experts in the GenAI Era

The Democratization of Ideation

In many ways, GenAI is the great equalizer.

Doshi and Hauser (2024) found that low-creativity participants improved their storytelling by 11% when given access to AI ideas. Not only did their performance increase, but the creative gap between novices and high performers virtually disappeared. AI raised the floor without lowering the ceiling.

This has profound implications for innovation. Students, junior employees, or people outside traditional innovation roles can now participate meaningfully in ideation. As Girotra and Meincke’s work suggests, with a few examples and a well-engineered prompt, anyone can contribute viable, high-quality ideas.

Art mirrors this trend. In AI-assisted haiku creation, collaborative efforts between humans and machines consistently outperformed both pure AI and pure human poems in aesthetic evaluations. GenAI helps amplify latent creativity, especially for those who lack formal training or confidence.

In short, GenAI levels the playing field.

The Expert Paradox: When Experience Gets in the Way

Ironically, experienced professionals don’t always benefit from GenAI — and in some cases, it may undermine their performance.

A striking example comes from a study by Eisenreich et al. (2024). When experts were shown AI-generated ideas for inspiration, they performed worse than either “pure” AI or “pure” human ideators. Why? The explanation seems to be anchoring — AI outputs may constrain creative thinking rather than catalyze it among seasoned minds.

This insight challenges the assumption that more expertise means better outcomes when using AI tools. Instead, it suggests a new skill is required: the ability to effectively collaborate with AI via guiding curating, and edited, but without being creatively boxed in.

Artists face the same challenge. In visual domains, Zhou and Lee (2024) found that those who simply plugged ideas into AI tools produced more generic work. But artists who curated and refined AI outputs saw the biggest boosts in evaluations and audience engagement.

The future expert isn’t just a creator. They’re a creative director, orchestrating a human-machine ensemble to push boundaries rather than settle into comfortable patterns.

Conclusion: From Brainstorming to Brainhacking

We are witnessing a historic shift — not just in how ideas are generated, but in who gets to generate them and what those ideas look like.

GenAI tools have redefined the ideation process. They produce more, faster, and often better. They empower novices, disrupt experts, and challenge our deepest assumptions about creativity. Yet they also introduce risks: homogenization, bias, and the temptation to outsource too much of our thinking to machines.

The challenge isn’t to resist GenAI, but to use it wisely. To know when to prompt and when to pause. To explore widely, then filter ruthlessly. To let GenAI flood the canvas, but retain the brush.

So the next time you need a breakthrough idea, don’t just think outside the box. Ask your favorite bot what it thinks the box should be made of.

Posted in AI, Creativity, Innovation | Tagged , , , , , , , , , | Leave a comment

The End of the Crowd? (Why AI Won’t Fully Replace Human Crowdsourcing — Yet)

Posted on May 12, 2025 by Eugene Ivanov

AI has already claimed its seat at the innovation table — and it didn’t even knock. It barged in, armed with large language models (LLMs) like GPT-4, reshaping how companies ideate, prototype, and solve problems. 

With astonishing speed and minimal cost, these tools are outperforming humans in tasks ranging from code generation to business model design. So, here’s the billion-dollar question: if AI can already outperform human crowds in many areas, is traditional crowdsourcing about to die?

A compelling study by Boussioux et al. (2024), titled “The Crowdless Future? Generative AI and Creative Problem Solving,” puts this debate into sharp focus. Their experiment pitted human-generated business ideas against those created using a human-AI hybrid approach. The results? AI-assisted solutions, especially when guided through strategically refined prompts, scored significantly higher in value, including financial and environmental impact, and overall quality. And they came with a price tag of just $27 compared to over $2,500 for the human-only submissions.

Translation? AI isn’t just good at creative problem-solving. It’s lean, scalable, and often better than the crowd, at least when measured by implementation potential and perceived value.

But if AI is that efficient, why aren’t we declaring the death of crowdsourcing right now?

While AI may outpace us humans in cost and consistency, there are at least four powerful reasons why traditional human crowdsourcing is far from obsolete.

Novelty: The Spark of the Unexpected

Boussioux et al. found that human-generated ideas consistently ranked higher in novelty, especially at the upper end of the scale. In other words, when you’re looking for that one-in-a-million idea — the weird, wild, breakthrough concept that no dataset can predict — humans may still have the edge.

AI models, no matter how advanced, are trained on what has been, not what could be. Their “creativity” is fundamentally synthetic — it’s a remix of the past. Human crowds, on the other hand, bring serendipity, fringe thinking, and unpredictable combinations. And in innovation, sometimes it’s one crazy idea, not a dozen “good” ones, that changes everything.

Ownership: Who Gets the Credit (and the IP)?

With AI-generated content, the question of intellectual property is still a legal and ethical minefield. If an LLM produces a groundbreaking idea based on prompts from your team, who owns the output? Your team? The model’s creators? The crowd of internet texts that the model was trained on?

Crowdsourcing sidesteps this ambiguity. A human contributor generates a breakthrough idea and signs an agreement transferring all IP rights to this idea to the crowdsourcing campaign sponsor in exchange for a reward, all in a legally transparent and unambiguous way. For organizations wary of future legal headaches, sticking with human solvers may feel like a safer bet, at least until AI governance frameworks catch up.

Marketing Value: Crowdsourcing as Innovation Theater

Let’s be honest: not all crowdsourcing is about getting the best ideas. Sometimes, it’s about signaling. When a company launches an open innovation contest — say, “Reimagine the Future of Food” — it’s making a statement: We’re listening to our customers. We’re cutting-edge. We’re engaged. Investors love this!

An AI prompt doesn’t generate press releases, Instagram buzz, or goodwill. But a vibrant campaign with real people submitting ideas does. For companies looking to boost their image as forward-thinking and innovative, the crowd still offers a potent narrative tool.

Community: It’s Not Just About the Ideas

Crowdsourcing doesn’t just produce solutions — it builds communities. When done right, it creates a network of passionate participants who care about a problem, become brand advocates, and sometimes even co-founders of spinoff ventures.

AI, by contrast, is transactional. It doesn’t care. It doesn’t get excited. It won’t show up at your hackathon or promote your brand on social media. That human energy — the sense of being part of something bigger — is still irreplaceable.

So, will AI replace crowdsourcing?

In many ways, it already has — for tasks where speed, scale, and strategic value matter most. But for organizations chasing radical novelty, craving emotional connection, or navigating uncertain legal waters, the human crowd still has a job to do.

Maybe the future isn’t crowdless — it’s crowdsmart. A hybrid world where AI augments, not replaces, the wisdom of the crowd. Where LLMs help us sift, refine, and accelerate, but humans still supply the spark.

In the end, it’s not AI vs. the crowd. It’s AI + the crowd. And when those two forces align, innovation doesn’t just scale — it soars.

Bold claim? Perhaps. But when the sparks fly from both silicon and soul, that’s when real innovation begins.

Posted in AI, Crowdsourcing, Innovation | Tagged , , , , | Leave a comment

Knowing Where You’re Going and Who’s Driving: How AI Is (and Isn’t) Reshaping Human Work

Posted on March 25, 2025 by Eugene Ivanov

Integrating AI into business practice has gone from a fringe conversation to a boardroom imperative. From productivity gains to fears of de-skilling, the debate is divisive—some see AI as a game-changer for human potential; others worry it’s a slippery slope toward dependence and displacement. 

Negative sentiments notwithstanding, the real question is not whether to use AI, but how to use it wisely. Four cutting-edge studies provide a nuanced view of this evolving frontier, shedding light on when AI helps, when it hinders, and how it may redefine not just work—but workers and teams themselves.

Inside the Frontier: When AI Knows What It’s Doing

The “jagged technological frontier” isn’t just a catchy metaphor—it’s the heart of a massive field experiment run with 758 consultants at the Boston Consulting Group. Researchers introduced GPT-4 to professionals tasked with solving complex business problems. 

The key insight? AI is only effective when operating within its capabilities—or “inside the frontier.” These are tasks that AI can complete reliably: structured analysis, clear communication, or ideation based on known patterns. “Outside the frontier” lies the domain of ambiguity, tacit knowledge, and judgment—and it’s here where AI stumbles and sometimes misleads.

The study’s surprising twist was who benefited the most from using AI tools. It wasn’t the top performers, but the consultants with below-average baseline performance. For these individuals, AI acted as an accelerant—boosting quality by over 40% and productivity by 25%. In contrast, for tasks outside the AI’s comfort zone, consultants with AI were 19% less likely to deliver the correct solution. These findings don’t argue against AI—they reveal its shape. Like any tool, AI is powerful only when used in the right context. Success comes from recognizing where AI’s frontier lies and then adapting accordingly.

Too Smart to Help? When Better AI Backfires

What happens when AI becomes too competent? 

In a striking counterpoint to exuberant techno-optimism, a 2022 work by Dell’Acqua earlier explores a phenomenon the author dubbed “falling asleep at the wheel.” In a field experiment with 181 professional recruiters, participants evaluated resumes with AI assistance. But this time, the quality of the AI tool varied—some recruiters received high-accuracy recommendations, others received low-accuracy ones.

Counterintuitively, the recruiters using lower-performing AI tools made better decisions. They were more engaged, spent more time reviewing resumes, and were more likely to challenge AI suggestions. Meanwhile, highly accurate AI caused the human effort to drop. Recruiters deferred too quickly to machine judgment and became less accurate in their assessments.

This wasn’t a fluke—it was particularly true for experienced professionals, whose own skills were diluted by over-reliance on the algorithm. 

The takeaway is clear: high-quality AI can displace rather than augment human expertise. In such settings, algorithmic excellence may seduce users into disengagement, suppressing their cognitive muscle memory. Maximizing joint performance may sometimes require less powerful AI—at least when keeping humans in the loop is critical.

Smarter Isn’t Always Better—But Sometimes It Is

Otis and colleagues offer a compelling twist to this narrative. In a randomized trial involving 640 Kenyan entrepreneurs, participants received business advice either from a traditional guidebook or via a GPT-4-based AI mentor on WhatsApp. Unlike the recruiter study, this AI tool helped top performers—boosting revenue and profits by over 20%. But it harmed low performers, who saw their performance dropping by about 10%.

Why this contradiction? It comes down to task selection and user discretion. Entrepreneurs had autonomy in when and how to use the AI, and high performers asked better questions on more manageable tasks. In contrast, low performers sought help on complex, ill-structured problems—those outside the AI’s frontier—leading to bad advice and worse outcomes.

This study makes more nuanced the notion that better AI leads to disengagement. It shows that it’s human judgment about what AI can and cannot do that is the real driver of success. When users are savvy about AI’s limitations, even powerful systems can be transformative. When they’re not, AI becomes a mirage—confidence without clarity.

Teaming Up with the Machine: A New Era of Collaboration

If the first three studies examined AI as a co-pilot for individuals, the just-published experiment conducted by a Harvard/Wharton team reimagines AI as a collaborator for entire teams. Conducted with 776 professionals at Procter & Gamble, the study asked: can AI fill the collaborative roles typically occupied by humans?

Participants were randomly assigned to four groups: individuals working solo, human teams of two, individuals with AI, and human teams with AI. All tackled real product development challenges. The results were eye-opening: individuals with AI matched the output of human teams. Even more striking, teams with AI outperformed all others, including human-only teams.

AI’s impact wasn’t just in better performance: it flattened functional silos. Without AI, participants generated ideas aligned with their functional background: R&D workers generated more technical proposals, and commercial workers more business-oriented. With AI, all produced more balanced solutions, regardless of background. Emotional benefits were evident too—users reported more positive feelings and less frustration when working with AI.

The implication of this study is profound: AI isn’t just a tool; it’s evolving into a cybernetic teammate, one that enhances creativity, bridges knowledge gaps, and even mimics the social glue of teamwork. (Who could predict this even a couple of years ago?) 

This shift could redefine how we structure teams, allocate expertise, and manage work across the enterprise. The age of the solitary “AI-enhanced worker” is giving way to something richer—and potentially more disruptive.

The AI Edge Depends on the Human Hand

Across four major field studies, a clear pattern emerges. AI can supercharge performance—but only when we understand how, when, and who should use it. It’s not the intelligence of the algorithm that matters most; it’s the alignment between task, user, and tool.

GPT-4 boosted underperformers—if tasks were within their skill set. High-quality AI backfired—if users relied on it blindly. Entrepreneurial outcomes varied—based on users’ understanding of AI’s strengths and limits. And now, AI isn’t just augmenting individuals—it’s enhancing teams.

As businesses race to adopt generative AI, the lesson is both simple and sobering: AI is only as good as the people who know how to use it. And isn’t it a case for all other tools?

So, are we ready to treat AI not just as a tool, but as a teammate? As a manager? Feel free to scratch out the last question.

Posted in AI, Creativity | Tagged , , , , , , , , , | Leave a comment

In Silico Creativity. Part 3. AI and Creative Art

Posted on March 4, 2025 by Eugene Ivanov

In Part 1 and Part 2 of this three-part series, I reviewed what is known about AI’s ability to generate poetry and music. This article is about what AI can do for creative art. 

As AI systems like DALL-E, Midjourney, and Stable Diffusion aggressively penetrate the art world, questions arise regarding how AI-generated works are perceived, valued, and integrated into human creative workflows. Two studies—one by Bellaiche et al. (2023) and another by Zhou & Lee (2024)—offer key insights into this debate, exploring AI’s role in augmenting human creativity and highlighting the biases influencing our appreciation of AI-generated art.

AI? No, please!

Bellaiche et al. (2023) investigate whether people prefer human-created artworks over those made by AI and, if so, why. Through a series of experiments, they find that individuals tend to rate artworks labeled as “human-created” more positively than those labeled “AI-created,” even when all images were, by the study design, produced by AI. The study shows that people associate greater meaning, effort, and emotional impact with human-made works—and this contributes to higher aesthetic evaluations.

Interestingly, participants with positive attitudes toward AI exhibited reduced bias against AI-labeled artwork. Additionally, people who scored lower on cognitive reflection tests were more likely to rate human-labeled art as more beautiful. That suggests that AI-created artwork is subject to top-down biases rather than bottom-up sensory judgments. 

These findings have important implications for AI-generated content in creative industries. While AI can produce high-quality artwork indistinguishable from human-made pieces, public perception remains a significant barrier to AI’s acceptance in creative fields. As AI-generated art becomes more common, overcoming biases and fostering appreciation for AI as a creative tool will be crucial.

AI? Yes, please!

While Bellaiche et al. focus on biases against AI-generated art, Zhou & Lee (2024) explore how AI enhances human creative productivity. Analyzing a dataset of over four million artworks, their research shows that AI can help art creators: AI-assisted artists experience a 25% boost in creative productivity and a 50% increase in positive evaluations of their work by peers.

However, the study also uncovers a paradox: while peak creativity—measured as content novelty—increases over time, average novelty declines. This suggests that while AI enables some artists to push creative boundaries and produce exceptionally novel artifacts, many others relying on AI’s capabilities begin producing aesthetically pleasing but less original work. 

The authors introduce the concept of “generative synesthesia,” describing the harmonious blending of human ideation and AI execution as a new form of creative workflow. This positions AI’s role not as a replacement for human creativity but as a tool that expands the creative process when used effectively.

The Evolving Landscape of Art Creation 

Together, these two studies offer a nuanced perspective on the role of AI in artistic creation. While AI can meaningfully enhance artistic output, psychological perceptions still favor human-made art. In other words, AI’s impact on creativity depends on how artists engage with AI tools and how audiences perceive the resulting work.

As AI-generated art will continue to proliferate—and there must be no doubt about that—it will be essential to address these biases and develop a more comprehensive understanding of creativity in the age of AI. Will society embrace AI-assisted art as a legitimate form of creativity, or will human authorship remain the gold standard? 

One thing remains clear though: the definition of art and the role of the artist are evolving, and AI is at the center of this transformation.

Posted in AI, Creativity | Tagged , , , , , , | Leave a comment

In Silico Creativity. Part 2. AI and Music

Posted on March 1, 2025 by Eugene Ivanov

As Artificial Intelligence (AI) keeps conquering creative fields, from visual arts to music, the debates on whether AI can be truly “creative” show no signs of abating. But a much more practical question is, do listeners perceive AI-composed music differently from human-composed? 

Two recent studies, one by Shank et al. (2022) and another by Zlatkov et al. (2023), explore these questions and reveal interesting insights into how people judge AI-generated music.

Do We Like AI-Composed Music Less?

The study by Shank et al. (2022) investigated whether people like music less when they believe it was composed by AI. In a first set of experiments, participants listened to excerpts of classical and electronic music and rated how much they liked them while also guessing whether they were composed by a human or AI. The results showed that listeners were more likely to assume that electronic music was AI-composed and tended to like it less if they believed this was the case.

In the next set, the researchers directly manipulated the information given to listeners about composer identity. They found that for classical music, participants liked the excerpts less when they were told it was AI-generated. This suggests a clear bias against AI composers, particularly in genres, like classical music, that are traditionally associated with human emotional expression and creativity.

…Or We Don’t Care?

The study by Zlatkov et al. (2023) explored a similar question from a different angle. Their experiment involved 163 participants who listened to five human-composed and five AI-composed musical pieces. The participants were divided into two groups: one was told the correct composer identity, while the other was deceived. The researchers hypothesized that those who knew that a piece was AI-generated would rate it lower.

Surprisingly, the evidence didn’t support this hypothesis. Unlike previous findings, Zlatkov et al. found that listeners did not necessarily dislike music just because it was AI-generated. However, researchers acknowledged limitations in their study design as they didn’t explore the role of musical style, listener background, and other contextual factors in shaping perceptions of AI-composed music.

…And Should We Care?

Both studies provide yet another example of the complexity of human perception of AI creativity. While one suggests that people have an inherent bias against AI-generated music, particularly when it challenges traditional notions of musical craftsmanship, the other indicates that this bias may not be as universal as previously thought; instead, it depends on context and how AI music is introduced to listeners.

The larger point, however, is that AI-composed music is good enough to fool people into believing that it was human-generated.

So, as AI-generated music becomes increasingly sophisticated, it’s not its quality but rather human perception that will represent a major hurdle to its adoption. Whether AI compositions will ever be embraced on equal footing with human-created music will therefore depend not just on technical advancements but also on changing cultural attitudes toward creativity itself.

And let’s be honest. You’re listening to a piece of music that gets you. Does it matter who—or what—composed it?

Posted in AI, Creativity, Science & Technology | Tagged , , , , , | 1 Comment

In Silico Creativity. Part 1. LLMs and Poetry (and Short Stories)

Posted on February 25, 2025 by Eugene Ivanov

In my previous article, “In Silico Ideation,” I reviewed academic literature describing the application of LLM algorithms to generating new product ideas. Now, I want to review what is known about LLMs’ ability to generate other creative content. This article is about poetry (and short stories). 

Can You Tell Who Wrote That Poem?

If you think you can easily spot the difference between AI-generated and human-written poetry, think again. In a 2024 study by Porter and Machery, 1,634 participants were randomly assigned to evaluate poetry from 10 well-known poets and poems generated by ChatGPT-3.5 written in the style of each poet.

Guess what? The participants failed to tell the difference between the two sets. Even more surprising, they were more likely to mistake AI-generated poems for human work than the other way around. Moreover, ChatGPT-3.5-generated poetry not only passed as human-written but was rated higher for overall quality, rhythm, and beauty compared to works by famous poets.

The researchers call this the “more-human-than-human” effect. When people like a poem, they tend to assume it must have been written by a human. This bias plays out consistently across experiments, regardless of participants’ experience with poetry.

However, there was a twist: when explicitly told that a poem was AI-generated, participants rated it lower than when told it was human-written, revealing persistent biases against machine creativity.

Enhancing Human Creativity

AI isn’t just creating content on its own—it’s also changing how humans create it. A 2024 study by Doshi and Hauser found that prior access to a pool of AI-generated “seed” ideas improved the novelty and usefulness of human-written short stories by 6.7% and 6.4% respectively. Stories inspired by AI prompts were also rated as more enjoyable and better written.

The most intriguing finding? AI appears to be a great equalizer. Writers with lower measured creative abilities saw improvements of up to 11% when using AI “seed” ideas, effectively closing the gap between them and their more naturally creative peers. 

The Collaboration Sweet Spot

It also appears that generating creative content is more effective when humans collaborate with LLMs rather than when either party works alone. A 2023 study by Hitsuwari and colleagues found that while AI-generated haiku and human-made haiku were rated equally beautiful, AI-generated haiku with human intervention received the highest beauty ratings. 

Again, participants couldn’t reliably distinguish between human and AI authors. Moreover, the higher the AI-generated haiku was rated, the more likely people were to believe it was human-made.

The Diversity Angle

There’s a potential downside to AI-induced creative enhancement. The Doshi and Hauser study found that AI-assisted stories showed higher similarity to one another and the AI-generated prompts. This suggests a reduction in the diversity of creative output, raising questions about AI’s role in fostering true originality over time.

Implications for the Future

These studies collectively point to several important implications:

1. Indistinguishable creation: The line between human and AI creativity is rapidly blurring, at least for shorter creative formats like poetry.

2. Democratization of creativity: AI tools can help level the playing field, potentially allowing those with less natural creative talent to produce work of similar quality to highly creative individuals.

3. The collaboration advantage: The highest quality creative output may come from human-AI partnerships rather than either working independently.

As AI continues to evolve, so will our understanding of creativity itself. Rather than seeing AI as a replacement for human creativity, these studies suggest we might be moving toward a future where AI becomes an extension of human creative capabilities—enhancing, equalizing, and potentially transforming how we create art.

The question isn’t whether AI can be creative, but how our collaboration with LLM systems will reshape the very notion of creativity itself. I’ll come back to this topic in my future articles.

Posted in AI, Creativity | Tagged , , , , , | 1 Comment