.jpg)
.jpg)
The surprising thing AI engineers will tell you if you let them
Ezra Klein, New York Times
last updated: Apr 22,2023
- +
- -
Among the many unique experiences of reporting on artificial
intelligence is this: In a young industry flooded with hype and money, person
after person tells me that they are desperate to be regulated, even if it slows
them down. In fact, especially if it slows them down.اضافة اعلان
What they tell me is obvious to anyone watching. Competition is forcing them to go too fast and cut too many corners. This technology is too important to be left to a race among Microsoft, Google, Meta, and a few other firms. But no company can slow down to a safe pace without risking irrelevancy. That is where governments come in — or so they hope.
A place to start is with the frameworks policymakers have already put forward to govern AI. The two major proposals, at least in the West, are the Blueprint for an AI Bill of Rights, which the White House put forward in 2022, and the Artificial Intelligence Act, which the European Commission proposed in 2021. Then, last week, China released its latest regulatory approach.
Use-based regulationLet us start with the European proposal, as it came first. The AIA tries to regulate AI systems according to how they are used. It is particularly concerned with high-risk uses, which include everything from overseeing critical infrastructure to grading papers to calculating credit scores to making hiring decisions. High-risk uses, in other words, are any use in which a person’s life or livelihood might depend on a decision made by a machine-learning algorithm.
The European Commission described this approach as “future-proof”, which proved to be predictably arrogant, as new AI systems have already thrown the bill’s clean definitions into chaos. Focusing on use cases is fine for narrow systems designed for a specific use, but it is a category error when it is applied to generalized systems. Models such as GPT-4 do not do any one thing except predict the next word in a sequence. You can use them to write code, pass the bar exam, draw up contracts, create political campaigns, plot market strategy, and power AI companions. In trying to regulate systems by use case, the AIA ends up saying very little about how to regulate the underlying model that is powering all these use cases.
Unintended consequences abound. The AIA mandates, for example, that in high-risk cases, “training, validation, and testing data sets shall be relevant, representative, free of errors, and complete”. But what the large language models are showing is that the most powerful systems are those trained on the largest data sets. Those sets cannot plausibly be free of error, and it is not clear what it would mean for them to be “representative”. There is a strong case to be made for data transparency, but I do not think Europe intends to deploy weaker, less capable systems across everything from exam grading to infrastructure.
The other problem with the use-case approach is that it treats AI as a technology that will, itself, respect boundaries. But its disrespect for boundaries is what most worries the people working on these systems. Imagine that “personal assistant” is rated as a low-risk use case and a hypothetical GPT-6 is deployed to power an absolutely fabulous personal assistant. The system gets tuned to be extremely good at interacting with humans and accomplishing a diverse set of goals in the real world. That is great until someone asks it to secure a restaurant reservation at the hottest place in town and the system decides that the only way to do it is to cause a disruption that leads one-third of that night’s diners to cancel their bookings.
Sounds like sci-fi? Sorry, but this kind of problem is sci-fact.
A broader blueprintThe White House’s Blueprint for an AI Bill of Rights is a more interesting proposal. But where the European Commission’s approach is much too tailored, the White House blueprint may well be too broad. No AI system today comes close to adhering to the framework, and it is not clear that any of them could.
The blueprint says that “automated systems should be developed with consultation from diverse communities, stakeholders, and domain experts to identify concerns, risks, and potential impacts of the system”. This is crucial, and it would be interesting to see the White House or Congress flesh out how much consultation is needed, what type is sufficient and how regulators will make sure the public’s wishes are actually followed.
Perhaps the most interesting of the blueprint’s proposals is that “you should be able to opt out from automated systems in favor of a human alternative, where appropriate”. In that sentence, the devil lurks in the definition of “appropriate”. But the underlying principle is worth considering. Should there be an opt-out from AI systems? Which ones? When is an opt-out clause a genuine choice, and at what point does it become merely an invitation to recede from society altogether, like saying you can choose not to use the internet or vehicular transport or banking services if you so choose.
Then there are China’s proposed new rules. I will not say much on these, except to note that they are much more restrictive than anything the US or Europe is imagining, which makes me very skeptical of arguments that we are in a race with China to develop advanced artificial intelligence. China seems perfectly willing to cripple the development of general AI so it can concentrate on systems that will more reliably serve state interests.
Five areas of focus for AI regulationAfter talking to a lot of people working on these problems and reading through a lot of policy papers imagining solutions, there are a few categories I would prioritize.
The first is the question — and it is a question — of interpretability. As I said above, it is not clear that interpretability is achievable. But without it, we will be turning more and more of our society over to algorithms we do not understand. If you told me you were building a next generation nuclear power plant but there was no way to get accurate readings on whether the reactor core was going to blow up, I would say you should not build it. Is AI like that power plant? I am not sure. But that is a question society should consider, not a question that should be decided by a few hundred technologists. At the very least, I think it is worth insisting that AI companies spend a good bit more time and money discovering whether this problem is solvable.
The second is security. For all the talk of an AI race with China, the easiest way for China — or any country for that matter, or even any hacker collective — to catch up on AI is to simply steal the work being done. Any firm building AI systems above a certain scale should be operating with hardened cybersecurity. It is ridiculous to block the export of advanced semiconductors to China but to simply hope that every 26-year-old engineer at OpenAI is following appropriate security measures.
The third is evaluations and audits. This is how models will be evaluated for everything from bias to the ability to scam people to the tendency to replicate themselves across the internet.
The fourth is liability. There is going to be a temptation to treat AI systems the way we treat social media platforms and exempt the companies that build them from the harms caused by those who use them. I believe that would be a mistake. The way to make AI systems safe is to give the companies that design the models a good reason to make them safe. Making them bear at least some liability for what their models do would encourage a lot more caution.
The fifth is, for lack of a better term, humanness. Do we want a world filled with AI systems that are designed to seem human in their interactions with humans? Because make no mistake: That is a design decision, not an emergent property of machine-learning code. AI systems can be tuned to return dull and caveat-filled answers, or they can be built to show off sparkling personalities and become enmeshed in the emotional lives of human beings.
I think the latter class of programs has the potential to do a lot of good as well as a lot of harm, so the conditions under which they operate should be thought through carefully. It might, for instance, make sense to place fairly tight limits on the kinds of personalities that can be built for AI systems that interact with children. I would also like to see very tight limits on any ability to make money by using AI companions to manipulate consumer behavior.
This is not meant to be an exhaustive list. Others will have different priorities and different views. And the good news is that new proposals are being released almost daily. The Future of Life Institute’s policy recommendations are strong, and I think the AI Objectives Institute’s focus on the human-run institutions that will design and own AI systems is critical. But one thing regulators should not fear is imperfect rules that slow a young industry. For once, much of that industry is desperate for someone to help slow it down.
Read more Opinion and Analysis
Jordan News
What they tell me is obvious to anyone watching. Competition is forcing them to go too fast and cut too many corners. This technology is too important to be left to a race among Microsoft, Google, Meta, and a few other firms. But no company can slow down to a safe pace without risking irrelevancy. That is where governments come in — or so they hope.
A place to start is with the frameworks policymakers have already put forward to govern AI. The two major proposals, at least in the West, are the Blueprint for an AI Bill of Rights, which the White House put forward in 2022, and the Artificial Intelligence Act, which the European Commission proposed in 2021. Then, last week, China released its latest regulatory approach.
Use-based regulationLet us start with the European proposal, as it came first. The AIA tries to regulate AI systems according to how they are used. It is particularly concerned with high-risk uses, which include everything from overseeing critical infrastructure to grading papers to calculating credit scores to making hiring decisions. High-risk uses, in other words, are any use in which a person’s life or livelihood might depend on a decision made by a machine-learning algorithm.
Should there be an opt-out from AI systems? Which ones? When is an opt-out clause a genuine choice, and at what point does it become merely an invitation to recede from society altogether
The European Commission described this approach as “future-proof”, which proved to be predictably arrogant, as new AI systems have already thrown the bill’s clean definitions into chaos. Focusing on use cases is fine for narrow systems designed for a specific use, but it is a category error when it is applied to generalized systems. Models such as GPT-4 do not do any one thing except predict the next word in a sequence. You can use them to write code, pass the bar exam, draw up contracts, create political campaigns, plot market strategy, and power AI companions. In trying to regulate systems by use case, the AIA ends up saying very little about how to regulate the underlying model that is powering all these use cases.
Unintended consequences abound. The AIA mandates, for example, that in high-risk cases, “training, validation, and testing data sets shall be relevant, representative, free of errors, and complete”. But what the large language models are showing is that the most powerful systems are those trained on the largest data sets. Those sets cannot plausibly be free of error, and it is not clear what it would mean for them to be “representative”. There is a strong case to be made for data transparency, but I do not think Europe intends to deploy weaker, less capable systems across everything from exam grading to infrastructure.
The other problem with the use-case approach is that it treats AI as a technology that will, itself, respect boundaries. But its disrespect for boundaries is what most worries the people working on these systems. Imagine that “personal assistant” is rated as a low-risk use case and a hypothetical GPT-6 is deployed to power an absolutely fabulous personal assistant. The system gets tuned to be extremely good at interacting with humans and accomplishing a diverse set of goals in the real world. That is great until someone asks it to secure a restaurant reservation at the hottest place in town and the system decides that the only way to do it is to cause a disruption that leads one-third of that night’s diners to cancel their bookings.
Sounds like sci-fi? Sorry, but this kind of problem is sci-fact.
A broader blueprintThe White House’s Blueprint for an AI Bill of Rights is a more interesting proposal. But where the European Commission’s approach is much too tailored, the White House blueprint may well be too broad. No AI system today comes close to adhering to the framework, and it is not clear that any of them could.
The blueprint says that “automated systems should be developed with consultation from diverse communities, stakeholders, and domain experts to identify concerns, risks, and potential impacts of the system”. This is crucial, and it would be interesting to see the White House or Congress flesh out how much consultation is needed, what type is sufficient and how regulators will make sure the public’s wishes are actually followed.
The easiest way for China — or any country for that matter, or even any hacker collective — to catch up on AI is to simply steal the work being done. Any firm building AI systems above a certain scale should be operating with hardened cybersecurity.
Perhaps the most interesting of the blueprint’s proposals is that “you should be able to opt out from automated systems in favor of a human alternative, where appropriate”. In that sentence, the devil lurks in the definition of “appropriate”. But the underlying principle is worth considering. Should there be an opt-out from AI systems? Which ones? When is an opt-out clause a genuine choice, and at what point does it become merely an invitation to recede from society altogether, like saying you can choose not to use the internet or vehicular transport or banking services if you so choose.
Then there are China’s proposed new rules. I will not say much on these, except to note that they are much more restrictive than anything the US or Europe is imagining, which makes me very skeptical of arguments that we are in a race with China to develop advanced artificial intelligence. China seems perfectly willing to cripple the development of general AI so it can concentrate on systems that will more reliably serve state interests.
Five areas of focus for AI regulationAfter talking to a lot of people working on these problems and reading through a lot of policy papers imagining solutions, there are a few categories I would prioritize.
The first is the question — and it is a question — of interpretability. As I said above, it is not clear that interpretability is achievable. But without it, we will be turning more and more of our society over to algorithms we do not understand. If you told me you were building a next generation nuclear power plant but there was no way to get accurate readings on whether the reactor core was going to blow up, I would say you should not build it. Is AI like that power plant? I am not sure. But that is a question society should consider, not a question that should be decided by a few hundred technologists. At the very least, I think it is worth insisting that AI companies spend a good bit more time and money discovering whether this problem is solvable.
The way to make AI systems safe is to give the companies that design the models a good reason to make them safe. Making them bear at least some liability for what their models do would encourage a lot more caution.
The second is security. For all the talk of an AI race with China, the easiest way for China — or any country for that matter, or even any hacker collective — to catch up on AI is to simply steal the work being done. Any firm building AI systems above a certain scale should be operating with hardened cybersecurity. It is ridiculous to block the export of advanced semiconductors to China but to simply hope that every 26-year-old engineer at OpenAI is following appropriate security measures.
The third is evaluations and audits. This is how models will be evaluated for everything from bias to the ability to scam people to the tendency to replicate themselves across the internet.
The fourth is liability. There is going to be a temptation to treat AI systems the way we treat social media platforms and exempt the companies that build them from the harms caused by those who use them. I believe that would be a mistake. The way to make AI systems safe is to give the companies that design the models a good reason to make them safe. Making them bear at least some liability for what their models do would encourage a lot more caution.
The fifth is, for lack of a better term, humanness. Do we want a world filled with AI systems that are designed to seem human in their interactions with humans? Because make no mistake: That is a design decision, not an emergent property of machine-learning code. AI systems can be tuned to return dull and caveat-filled answers, or they can be built to show off sparkling personalities and become enmeshed in the emotional lives of human beings.
One thing regulators should not fear is imperfect rules that slow a young industry. For once, much of that industry is desperate for someone to help slow it down.
I think the latter class of programs has the potential to do a lot of good as well as a lot of harm, so the conditions under which they operate should be thought through carefully. It might, for instance, make sense to place fairly tight limits on the kinds of personalities that can be built for AI systems that interact with children. I would also like to see very tight limits on any ability to make money by using AI companions to manipulate consumer behavior.
This is not meant to be an exhaustive list. Others will have different priorities and different views. And the good news is that new proposals are being released almost daily. The Future of Life Institute’s policy recommendations are strong, and I think the AI Objectives Institute’s focus on the human-run institutions that will design and own AI systems is critical. But one thing regulators should not fear is imperfect rules that slow a young industry. For once, much of that industry is desperate for someone to help slow it down.
Read more Opinion and Analysis
Jordan News
