This just feels like non-technical fear mongering. Frankly, the term “AI” is just way too overused for any of this to be useful - Autopilot, manufacturing robots, and ChatGPT are all distinct systems that have their own concerns, tradeoffs, regulatory issues, etc. and trying to lump them together reduces the capacity for discussion down to a single (not very useful, imo) take
editing for clarity: I’m for discussion of more regulation and caution, but conflating tons of disparate technologies still imo muddies the waters of public discussion
If you read the article, the concern is how those disparate technologies are converging.
You get the picture. Robots—“intelligent” and not—have been killing people for decades. And the development of more advanced artificial intelligence has only increased the potential for machines to cause harm. Self-driving cars are already on American streets, and robotic “dogs” are being used by law enforcement. Computerized systems are being given the capabilities to use tools, allowing them to directly affect the physical world. Why worry about the theoretical emergence of an all-powerful, superintelligent program when more immediate problems are at our doorstep? Regulation must push companies toward safe innovation and innovation in safety. We are not there yet.
I read the article, and stand by my statement - “AI” does not apply to self driving cars the same way as robotics use by law enforcement. These are two separate categories of problems where I don’t see how some unified frustration at AI or robotics applies.
Self driving cars have issues because the machine learning algorithms used to train them are not sufficient to navigate the complexities of roads, and there is no human fallback. (See: autopilot)
Robotics use by law enforcement has issues because it removes a human factor to enforcement, which has concerns of whether any deadly force is ever justified when used (does a suspect pose a danger to any officer if there is no human contact?), and worries of dehumanization exist here, as well as other factors like data collection. These aren’t even self driving mostly, from what I understand law enforcement remote pilots them.
these are separate problem spaces and aren’t deadly in the same ways, aren’t unattractive in the same ways, and should be treated and analyzed as distinct problems. by reducing to “AI” and “robots” you create a problem that makes sense only to the technically uninclined, and blurs any meaningful discussion about the precisions of each issue.
AI and robotics companies don’t want this to happen. OpenAI, for example, has reportedly fought to “water down” safety regulations and reduce AI-quality requirements. According to an article in Time, it lobbied European Union officials against classifying models like ChatGPT as “high risk,” which would have brought “stringent legal requirements including transparency, traceability, and human oversight.” The reasoning was supposedly that OpenAI did not intend to put its products to high-risk use—a logical twist akin to the Titanic owners lobbying that the ship should not be inspected for lifeboats on the principle that it was a “general purpose” vessel that also could sail in warm waters where there were no icebergs and people could float for days.
What would’ve been high risk? Well:
In one section of the White Paper OpenAI shared with European officials at the time, the company pushed back against a proposed amendment to the AI Act that would have classified generative AI systems such as ChatGPT and Dall-E as “high risk” if they generated text or imagery that could “falsely appear to a person to be human generated and authentic.”
That does make sense, considering ELIZA from the 60s would fit this description. It pretty much repeated what you wrote to it in a different style.
I don’t see how generative AI can be considered high risk when it’s literally just fancy keyboard autofill. If a doctor asks ChatGPT what the correct dose of medication for a patient is, it’s not ChatGPT which should be considered high risk but rather the doctor.
This just feels like non-technical fear mongering. Frankly, the term “AI” is just way too overused for any of this to be useful - Autopilot, manufacturing robots, and ChatGPT are all distinct systems that have their own concerns, tradeoffs, regulatory issues, etc. and trying to lump them together reduces the capacity for discussion down to a single (not very useful, imo) take
editing for clarity: I’m for discussion of more regulation and caution, but conflating tons of disparate technologies still imo muddies the waters of public discussion
If you read the article, the concern is how those disparate technologies are converging.
I read the article, and stand by my statement - “AI” does not apply to self driving cars the same way as robotics use by law enforcement. These are two separate categories of problems where I don’t see how some unified frustration at AI or robotics applies.
Self driving cars have issues because the machine learning algorithms used to train them are not sufficient to navigate the complexities of roads, and there is no human fallback. (See: autopilot)
Robotics use by law enforcement has issues because it removes a human factor to enforcement, which has concerns of whether any deadly force is ever justified when used (does a suspect pose a danger to any officer if there is no human contact?), and worries of dehumanization exist here, as well as other factors like data collection. These aren’t even self driving mostly, from what I understand law enforcement remote pilots them.
these are separate problem spaces and aren’t deadly in the same ways, aren’t unattractive in the same ways, and should be treated and analyzed as distinct problems. by reducing to “AI” and “robots” you create a problem that makes sense only to the technically uninclined, and blurs any meaningful discussion about the precisions of each issue.
What would’ve been high risk? Well:
That does make sense, considering ELIZA from the 60s would fit this description. It pretty much repeated what you wrote to it in a different style.
I don’t see how generative AI can be considered high risk when it’s literally just fancy keyboard autofill. If a doctor asks ChatGPT what the correct dose of medication for a patient is, it’s not ChatGPT which should be considered high risk but rather the doctor.
Isn’t that The Atlantic’s MO?