随着SpaceX、OpenAI和Anthropic计划推出重磅产品，这会使人工智能巨头更负责任吗？

A huge change is coming to the world’s booming artificial intelligence (AI) sector.

Starting with Elon Musk’s SpaceX, with OpenAI and Anthropic preparing to follow, all three private companies are set to sell shares of their stock to the general public for the first time. These are what’s known as initial public offerings (IPOs) .

SpaceX – the first of them to launch this Friday, June 12 – expects to raise $US75 billion from selling just 4% of the company’s shares.

Musk is already the world’s richest man, worth around US$800 billion. He owns around 42% of SpaceX now, plus options to buy more shares at a fraction of the US$135 a share price ordinary investors are being asked to pay. Given his existing wealth, after this Friday’s listing Musk looks likely to become the world’s first trillionaire.

Together, these three companies are valued at almost $US4 trillion and are expected to raise a record-breaking $US200 billion, despite well-founded concerns that big AI stocks are now hugely overvalued.

While most of the news coverage has focused on the money involved, there’s actually another side to these sales that could be a big deal in the longer run.

At a time when everyone from the Pope to people from all walks of life worldwide are concerned about AI’s growing role in our lives, these stock exchange listings have the potential to finally bring some extra transparency to the inner workings of the AI giants.

Why SpaceX, OpenAI and Anthropic matter to you

Once these companies list, hundreds of millions of investors around the world will be exposed to these companies. That could be directly, if you buy these stocks, or else through index funds, which hold shares on behalf of investors – including big retirement and superannuation funds.

Even for those who don’t consider themselves investors, these three share offerings could easily affect your savings too.

Here’s what we know about these IPOs so far.

SpaceX’s most recent June 3 filing amendment with the United States Securities and Exchange Commission added a notable new line. It said SpaceX “may issue a significant amount of equity in connection with future transactions”.

US business outlet Fortune is reading this as a signal for a possible future Tesla merger, bringing another of Musk’s companies into the fold. That could be the biggest merger in history.

Investors can sue over failures to disclose

Once publicly listed, the AI labs of SpaceX – xAI – as well as Anthropic and Open AI would be subject to public market scrutiny for the first time.

This would push these companies to disclose more AI risks than they have had to as private companies – or risk being sued for misleading investors.

US securities laws are among the most enforceable in the world. Under US law, investors can sue a company for securities fraud if it fails to disclose a risk that later materialises.

One regulation commonly used in securities fraud lawsuits is Rule 10b-5 under the Securities Exchange Act of 1934.

This has been successfully used in the past many times. For example, Bank of America paid US$2.43 billion settle a lawsuit related to its purchase of investment bank Merrill Lynch in 2008. Countrywide Financial paid US$600 million for failing to disclose the mounting risks of its subprime mortgage business.

Only last month, the International Monetary Fund warned “financial stability risks mount as artificial intelligence fuels cyberattacks”, pointing out:

Anthropic’s recent controlled release of its Claude Mythos Preview, an advanced AI model with exceptional cyber capabilities, underscored how quickly risks are increasing […] This foreshadows how fast‑moving, AI‑driven cyber risks could destabilize the financial system if not managed carefully.

There are good reasons to be concerned about the increasing dominance of tech companies and what happens to economies around the world if the AI share bubble bursts.

Having more of the biggest AI companies forced into greater disclosure would offer one silver lining amid those AI fears.

What difference could public disclosure make?

Just as an example, let’s suppose Anthropic accidentally leaked its Claude Mythos source code (like a leak that actually happened earlier this year) . Then let’s say North Korean hackers used that code to hack into US government systems.

If that happened when Anthropic was a public company, its share price would very likely fall in response.

Investors could then sue Anthropic for failing to disclose the risk of code leak, which later caused the share price to fall.

This mechanism has its limitations: it only works if AI harms are eventually reflected in stock prices of Anthropic.

In other words, the mechanism only protects the general public from AI risks indirectly – though protecting Anthropic’s investors first.

How much more accountability should we expect?

The market is meant to incorporate all public information to arrive at the fair price of a public company.

In doing so, market listings should make it easier for investors to police AI safety. After all, it’s in investors’ interests to not drive humanity to the verge of collapse.

But is the market delivering on this function so far with AI?

So far, you’d have to say it’s not. For instance, the world’s second-largest stock exchange, the New York-based Nasdaq, controversially changed its own rules for SpaceX to join its Nasdaq 100 index after just 15 trading days, not the usual three months.

But perhaps there is still hope that investors’ own desire to survive AI will make them push companies to manage AI’s risks more responsibly.

Will it be enough? Probably not on its own. The risks most people worry about with AI – diffuse, slow-moving, hard to pin to a single quarter – may never register clearly in an earnings report.

But more disclosure is better than less. And more disclosure is exactly what these listings will finally force.

Marta Khomyn does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.