px_eliezer_ Some of your links actually reinforce my argument.
First off, it’s normal for lawsuits to target everything and everyone. A man in the US recently electrocuted himself to death trying to break into an ATM machine and now his estate is suing the manufacturer of the ATM machine because they designed their ATM machine in such a way that a criminal could injure themselves while trying to steal money. No joke, google it.
With that being said, a hospital has its own responsibilities when selecting and overseeing medical professionals, just as a business has responsibilities when selecting and overseeing technology providers.
The existence of another potentially liable party doesn’t eliminate those responsibilities.
You’re conflating two different issues - an organization’s responsibility for the services it provides and the possibility that other parties may also share liability when something goes wrong.
My argument isn’t that an AI developer can never be held responsible. My argument is that a company cannot absolve itself of responsibility for its own operations simply because it outsourced some of those operations to an AI provider and we as a society should not allow that to happen.
This isn’t a new concept. In fact, American banking regulators explicitly state that using third-party service providers does not diminish a bank’s responsibility to comply with its legal and regulatory obligations. Canadian banking regulators take the same position.
If a bank negligently implements an AI system that causes financial losses or compromises customer information, the bank doesn’t get to shrug its shoulders and blame the AI company. It’s responsible for properly evaluating, implementing, securing and overseeing the systems it uses.
Now, if the AI vendor supplied a defective product, misrepresented its capabilities or otherwise acted negligently, that’s a separate issue. The vendor may also be responsible to the recipient of those services. One doesn’t cancel out the other.
And the surgeon analogy actually supports my point. A hospital has responsibilities concerning the qualifications and oversight of the people it allows to practice there. The surgeon’s potential liability doesn’t somehow eliminate the hospital’s independent obligations.
The same principle applies to software, cloud services, cybersecurity and AI. Organizations have been outsourcing technology for decades, and the responsibility to manage the associated risks didn’t suddenly disappear when AI entered the picture.
If the organization implemented the technology negligently, hold the organization accountable. If the technology provider was negligent, hold the provider accountable. If both were negligent, hold both accountable.
But <B>blaming AI developers for every failure involving AI is no more reasonable than automatically blaming Microsoft whenever a business experiences a Windows-related outage.</B>