BY DAVID SCHONTHAL, DIRECTOR OF ENTREPRENEURSHIP
KELLOGG SCHOOL OF MANAGEMENT@DAVIDSCHONTHAL
When it comes to the possibility that generative A.I. platforms -- the next generation of GPT-4, Midjourney, and the like -- could engage in some form of entrepreneurship without human involvement, many business-builders would say, "Not any time soon" or "Even if they could, it won't affect my business."
Both of those are naive responses at best. Wishful thinking.
There's no reason to think a GPT-5 or 6 couldn't soon be capable of building, launching, and growing a viable business, especially in the broad tech space.
Why am I so confident about this possibility? Because entrepreneurship, at its core, is algorithmic, based on the underlying sequence of sense-act-learn: It's about sensing or identifying patterns in terms of meaningful, addressable gaps in consumer needs, acting by iteratively creating and testing solutions to fill those gaps, taking the product or service that emerges from that process to market, and learning at each step to revisit and refine the offering.
We can all probably agree that generative A.I. has got the algorithm thing down better than we humans do -- and as a bonus, A.I.-based technologies can theoretically perform each of those business-building steps while considering more novel creative possibilities than we could and avoiding bias.
If you're still with me, imagine BB-1 (Business-Builder 1), a hypothetical entrepreneurial A.I. platform, going through the typical lean startup process. Say it's been tasked with exploring -- or decided on its own to explore -- development of a better digital workplace-collaboration resource, something that pushes the limits of and competes with market-leading products like Slack.
BB-1 could start by using natural language processing (NLP) to scan the entirety of the global public internet and identify and synthesize unmet needs related to resources in this space, such as by examining online comments about current collaboration resources, especially content pointing to what's missing from or frustrating about today's offerings. Next, the platform could study the many terabytes of code written by the 100 million-plus developers on GitHub to code a minimum viable product (MVP) that meets the most common unmet needs it has identified. The basic software product it creates could then be tested iteratively by recruiting early users through automated, digital usability studies, A/B tests, and other statistics-based techniques to refine the offering before going to market.
All of that could happen in a tiny fraction of the months or likely years it would take a team of humans to create the same offering -- and with every iterative cycle, the offer would materially improve. That's why I think generative A.I. could soon represent the leanest of lean startups, with minimal human intervention.
A bit spooky, right?
Well, it doesn't have to be. At least not yet. For now, let's assume generative A.I. isn't in a garage somewhere building a startup that will be the next big thing in digital. Indeed, many argue A.I. simply isn't ready for prime time even when it comes to unsupervised decision-making, because it lacks empathy, originality, and ethics. And even if entrepreneurial A.I. is an inevitability, the most important thing for now is to be aware of the possibility and to understand what it means for you as a current or aspiring business-builder.
Here are the main takeaways to consider:
Don't seek refuge in denial.
As I noted earlier, we humans are overly skilled at the head-in-the-sand strategy: "It's not going to happen. And even if it does, it's not going to affect me." That's wrongheaded and dangerous, as countless iconic business failures suggest (see Kodak and Blockbuster, among others). Assume A.I.-launched businesses will happen at some point in the not-too-distant future and prepare accordingly, using the ideas below.
Embrace augmentation.
For now, A.I. remains firmly in the augmentation category, serving to enhance and supplement decision-making, operations, and service across domains from consumer goods to sports. So there's every reason to use A.I. to your advantage, whatever your business. Specifically, find a way to leverage A.I. to sense, act, and learn, whether using machine learning to mine data for unmet needs in your target population, develop more creative products, or evaluate your current offerings without emotion or bias. Indeed, using A.I. could enable you to streamline early-stage operations to just you (the one with the idea), a data scientist, and a developer, for a super-lean approach.
Be inspired by A.I.
Most importantly, take cues from how A.I. might build or grow a business to work on your own -- rather than denying the superiority of A.I. algorithms or believing humans will always "win" just because we're, well, human. The best entrepreneurs, in fact, behave a bit like generative A.I.: they sense, act, and learn, with a systematic, rigorous approach at each step, for better identification of opportunity, crafting of ideas and solutions, and improvement of offerings. For now, A.I. does some of that (OK, a lot of it) better than we do, but it's still subject to flaws such as overconfidence, bias, originality, and even hallucinations. So take inspiration from A.I.'s approach to the core entrepreneurial process, while avoiding the very human-seeming mistakes today's A.I. makes.
When we look back at A.I.'s advancement, we'll recognize that it happened faster than most of us ever thought possible. The ideas here will help you become the best entrepreneur you can be, so you'll be prepared not only to survive what's ahead in our digital future but to make the most of it.
Comments