This is either a smart step toward making public markets less of a private club, or it’s the start of a paperwork arms race that rewards the fastest clicker, not the most honest company.
OpenAI partnering with Cooley to launch “GO Public” sounds simple: an AI tool that can draft an S‑1 filing in minutes instead of days. The S‑1 is the big registration document companies file with the SEC when they want to go public. In plain terms, it’s the story a company tells investors about what it is, how it makes money, and what could go wrong. And it’s not just a story—it’s a legal document people rely on.
Based on what’s been shared publicly, the pitch here is speed. Fewer hours of lawyers and staff grinding through drafts. Less waiting. Less cost. If you’ve ever watched a company drown in “we need this in the right format” work, you can see why this feels appealing.
But I don’t think speed is the win people think it is. Not by itself.
Because the hard part of an S‑1 is not typing. The hard part is deciding what you’re willing to admit in public. The hard part is the uncomfortable conversations: “Do we disclose that customer churn is rising?” “Do we call this revenue ‘recurring’?” “Do we explain that one big customer could walk and wreck us?” That’s the stuff that slows teams down—and it should. Slowness is often the point when you’re asking the market to trust you with billions.
If GO Public really does turn days into minutes, what exactly is being compressed? Best case: it compresses formatting, repetition, and first drafts, and frees humans to focus on the true judgment calls. That’s the optimistic view, and it’s not crazy. A lot of legal work is template-driven. A lot of first drafts are basically “last time, but change the nouns.”
The darker view is that it compresses accountability. It gives founders and bankers a shiny new way to say, “We’ve got a draft,” before anyone has actually done the uncomfortable thinking. A draft creates momentum. Momentum becomes a deadline. A deadline becomes pressure. And pressure is when weak disclosures slip through, not always out of malice, but because everyone is tired and the document looks “done.”
Imagine you’re a CFO at a fast-growing company and your board wants an IPO window. You’re already juggling audits, roadshow prep, product issues, and hiring. Someone drops a tool on your lap that spits out a clean-looking S‑1 in minutes. Are you going to slow down and pick apart every risk sentence? Or are you going to treat it like a good-enough starting point and trust the lawyers to catch the rest later?
Now imagine you’re a junior lawyer or a compliance person. Your job is to be the annoying one who says, “Hold on.” But the new expectation becomes: “The tool got us 80% there. Why are you slowing us down?” That’s not a small cultural shift. That’s how people get trained out of carefulness.
There’s also a fairness argument here, and I actually sympathize with it. The current IPO process is expensive and intimidating. If AI tools lower the barrier, maybe more companies can enter public markets without burning a fortune on drafting and redrafting. Maybe smaller companies stop being forced into private funding paths that concentrate power with a few players. If this tool makes the process more accessible, that’s a real benefit.
But accessibility cuts both ways. Making it easier to produce “IPO-ready” paperwork can also make it easier for shaky companies to look polished. Investors don’t buy a filing; they buy what the filing claims is true. A smoother document can create false comfort. And if everyone starts using similar AI drafting, filings may become more uniform, more generic, and less useful. When every risk section reads like it came from the same machine, the signal gets weaker.
There’s another tension people won’t like: if AI can draft these filings in minutes, regulators and investors will start expecting faster responses, more frequent updates, and more documentation. Speed doesn’t just reduce cost; it can increase volume. So instead of “great, less work,” you get “great, more work, but now it’s instant.” That’s how tools meant to save time end up raising the baseline.
And what happens when something goes wrong later? If a company’s disclosures are challenged, who owns the wording? The founders who signed? The lawyers who reviewed? The tool that drafted? In reality, responsibility will still land on humans. But humans will be tempted to treat AI text like a neutral starting point, as if it arrived without bias. It didn’t. Every tool has defaults. Defaults shape what gets emphasized, what gets softened, what sounds “normal.”
I’m not against using AI in legal workflows. I’m against treating legal writing like it’s mainly a typing problem. The real value of good counsel is judgment and pushback, not speed. If GO Public becomes a way to get to that judgment faster, great. If it becomes a way to avoid that judgment, it’s dangerous.
So here’s the question I can’t shake: does making S‑1 drafting faster mostly improve honesty by freeing time for deeper review, or mostly weaken honesty by turning a serious disclosure process into a rapid-content workflow?