No open-access location is known for this record yet.
No open-access location is known for this record yet.
In law and policy debates, there are many terms that get thrown around—big data, algorithms, artificial intelligence. One way to understand the technology is to think of the big data as the fuel, algorithms as the rockets; and artificial intelligence as the planet that computer scientists seek to reach.1 That is, to go in reverse order, the goal of artificial intelligence is to empower computers to replicate all the things that humans can do—see, hear, speak, even think. And how is that to be accomplished? By algorithms that use big data. Or, more precisely, by machine learning algorithms that use big data. Machine learning algorithms represent a newer generation of computer science. Old style algorithms are based on complete models with relatively few explanatory variables and contain a comprehensive set of if-then statements that give instructions to a computer. Machine learning algorithms are very different: based on an initial algorithm and the data inputs and the desired output, they do the work of generating what can be an extraordinarily complex, operating algorithm. To quote from the computer scientist Pedro Domingos: Every algorithm has an input and an output: the data goes into the computer, the algorithm does what it will with it, and out comes the result. Machine learning turns this around: in goes the data and the desired result and out comes the algorithm that turns one into the other. Learning algorithms—also known as learners—are algorithms that make other algorithms. With machine learning, computers write their own programs, so we don’t have to.2 The difficulty with machine learning, at least for the law, is that the actual content of that algorithm is often not fully known or knowable to the humans operating the code, not even to the scientific expert: State-of-the-art machine learning deploys far more complex models to learn about the relationship across hundreds or even thousands of variables. Model complexity can make it difficult to isolate the contribution of any particular variable to the result. . . [R]elatedly, the machine learning outputs are often nonintuitive—that is, they operated according to rules that are so complex, multi-faceted, and interrelated that they defy practical inspection, do not comport with any practical human belief about how the world works, or simply lie beyond human-scale reasoning. Even if data scientist can spell out the embedded rule, such rules may not tell a coherent story about the world as humans understand it, defeating conventional modes of explanation.3 Only recently has the law sought to address the new scientific and human reality. The novelty that the emerging legal frameworks seek to capture is the ability of machines to do what only a few decades ago most people thought only humans could do. So far, in the United States, no one legal definition of the technology and the policy problem has emerged as dominant. As the Administrative Conference of the United States puts it: There is no universally accepted definition of “artificial intelligence,” and the rapid state of evolution in the field, as well as the proliferation of use cases, makes coalescing around any such definition difficult. . . Generally speaking, AI systems tend to have characteristics such as the ability to learn to solve complex problems, make predictions, or undertake tasks that heretofore have relied on human decision making or intervention. There are many illustrative examples of AI that can help frame the issue for the purpose of this Statement. They include, but are not limited to, AI assistants, computer vision systems, biomedical research, unmanned vehicle systems, advanced game-playing software, and facial recognition systems as well as application of AI in both information technology and operational technology.4 There are a couple of definitions that have gained currency in the law. The first is contained in the John S. McCain National Defense Authorization Act for Fiscal Year 2019,5 and is used in the AI in Government Act of 2020,6 Executive Order 13960,7 and Office of Management and Budget, M-21-06.8 It reads as follows: (1) Any artificial system that performs tasks under varying and unpredictable circumstances without significant human or that can learn from and to data artificial system in computer software, or that tasks learning, or artificial system to think or a and set of machine learning, that is to a artificial system to an or that learning, decision and The definition is contained in the National Act of It The artificial intelligence a system that for a set of make predictions, or or intelligence systems use machine and inputs and such into models in an and use to for information or there are definitions that have as of for on facial and One is the and is at of algorithms on a one from machine learning or other artificial intelligence that information or other data for the purpose of the or that a set of information is to, or from a of an the of the of or any other of decision content or content In with the state of this the terms AI and algorithms as is often the the law in the and computer of algorithm can to style computer or machine There are a couple of other to in to the the to not to the state law the of the of the conventional the of and of the and legal frameworks that in has the of of algorithms does it to that of a algorithm is an algorithm is used by it or data and algorithms have in at least the that the Act to the of data contained in the computer systems of the complex algorithms and big data with AI are to the In of data for of and In the machine learning in the is for the from to that of the or with machine learning technology to out a of used machine learning for more one in use across the use cases, or fully the in the or or all the use algorithms the most in their decision the law out as the of machine that are and algorithms used across or or There are no comprehensive data on the use of AI at the state and one that in the is and is a state and in the United is a used by to and to to The system algorithms that on a to a as well as and other of data, the a to the on the that can be used to a decision at the the has their a decision on to there policy to in of their the use of algorithms has and has for and all of law is a big of the AI that have by the have to the big data and to of One that to in the use by and of data to with in the of has and and on their to one of the and a big data data the and other of turns around and that information to of is a to and the has used the data, to this data and to the and under from in the and they have the data. recognition technology is of how law has on there has not a at least at the The big data of with with machine learning algorithms, have the of facial recognition technology can be used for the of of to the have that to use an to a facial recognition The has to be at the of people of as to of facial recognition has for human in a the the to the use of facial recognition has and a a but so far no has the state there are a of and in a law by has to use facial recognition but to and at the and a couple of other have the use of facial recognition by the and other a on decision making in The United States has a on The the in in Executive Order of Office of Management and Budget, for of of AI in Executive Order of AI in Government Act of National Act of To Executive and are the only legal that to set comprehensive for speaking, the Executive the AI in the and and AI in the The Executive Order is at and to AI the and in the and The Executive Order is on the of this It is at of AI that can of their It all with the of and on of Executive Order It that the of Executive is In on AI law is the of this the to the of AI use and is not and this will the set in The are and in The the AI in Government Act of a in the the AI of is with of AI in the The the National Act of a for AI in all The of this is the National of the Office of and the on decision making to an data Executive Order to and fully data, and to the of such for AI and with and the on the of the National Office are the and are to data at the and other the what of systems are or advanced learning and learning decision making or or a of Any of AI that will and at the as is are for the of algorithms in The most of the for AI in is in Executive Order It the intelligence to the of the United States and the of of all In with Executive Order of in and have the of AI to their and of the use of and of and decision making by the other the of and by can from the use of or data to a an The that have as of the the of use of data that is of the data is not as or and it is not to an the law for of or data by The Government as of the for Act of that make data by As in a on the of the Act by the the to make data and for the and the the for Act of of the and Government Act of Government to their information as data making data to the by The and on data policy the Office of Management and as an The Government Act that and comprehensive data and with such as and the to understand how and use data. The to a that have an data that data for on the In to the Government Act to by and on in their data with the to a data of and with and the National and to a of and to data or to artificial intelligence in There are a of for AI in In the examples of law the technology is and by the and it is by the In other cases, it has In other cases, it is the of and the To give a of of the modes of the is the use that of the AI by from the the the of with and to algorithms into There are many in the to both and in policy at the and The National Office is in the Office of and is the on AI and as a of for and the The has an AI of the of AI across the The National of and has research, and policy to There are The is an illustrative AI of and Office Office of of National there for the of an algorithm with both and There are a couple of in to a have the big data for algorithms, but there are no for a algorithm In many of the the is does in decision making under the The legal at the of this do not contain a definition of It is to state law for into there are no state to algorithms speaking, has recently a law on facial It a state or to use facial recognition technology it an with the the and on the and based on facial recognition that legal be to human at the on the of the is to of the of the are at the data that are by and data and data are not only to the data but to the data of is and to algorithms more the the of of facial recognition in operational It on facial recognition data for of data is to the on the law rules both the use of algorithms and the of algorithms in Executive Order the for of AI in It and of and and and and to the of and AI in the The for AI are set in of Executive Order the law on algorithms and As there are Executive for the use of algorithms by Order and for the of and algorithms Order The contained in the Executive in the set out in to the of and is into the law significant algorithms and the of in a in of the of complexity of the That is, are only to and are to human the law both decision making and the law of in decision making in a or or or it of a few Executive Order is in this law law or all Executive Order is law that is on the Executive in in but they can be at will by the are the varying of of law, law, law, law, and in this the the United States to on AI Executive Order is not in and law be to under Executive Order this law a legal for data and it data and there are that are The Act of a of the there is for the of data on how their have a to they systems that use data the Act of the Act of give to or other of the law for the use of and in and by is in and state law the the To the has not the issue of is by the In a of in the the that there no of by conventional and the not the and of is from of and it is how the will with the the state there are that the use of and that the a facial recognition is one of the most used by and In the of the and data of is by the and Act of have in terms of and to and in the of on and As to of and it may very well the to It and in a the that AI may have on AI have the of by human the in that or that and in AI or be used in other that or to AI in with law, of and with to and by the AI application at as well as the AI application at issue may of or as to data the only for the of decision law, there are no to AI is as an of the of that can legal under law and law. use of AI may be to the of a rule, in the of the Administrative Act may In the of the use of AI may be to or any of from an AI system could be as a of the and the law a to a human in the be to of the this from the initial decision or only at the law, there is no such that the to the to a human in but the of is for this to have way into the or frameworks are to a human and frameworks such as the one by the the the and or There are no frameworks to a human and With to AI speaking, as there is by the National in the Office of and and the AI of There are all the The Administrative Conference of the United States how be to It is that AI systems be to and their There are of and of will be by the of be of both of in making about their AI of of AI systems can from a of the Government and In of AI systems to in a of can an in . a for AI systems the That is if a system or the circumstances in it is are to In and of the at one of or use may to in the To that of and of the data used by their AI systems, as well as how the systems use data. may to input from of the or to the that they will in their AI is the definition of in the of in decision Executive Order for the use of AI in at least of are to AI be the of AI systems be and and information on AI use be to of are in that the and of their AI are by and as and that human and are and the and use of that AI is used in a with and the for use of AI is The and use of as well as inputs and outputs of particular AI be well and as and to the be in information their use of AI to the and the to the and in with and with to the of and of law and other In on As the to of and there can be to be significant in the to AI and the of making AI to both and the that can be to be by the the it in of from of the information by the on and the of the AI The is in and In to the and can and in AI by to understand how an AI application and to understand the by AI a be in a that is for the to understand and may AI is in for if for about how the application human may be to the ability of human and other of the to make be that of AI could or human decision the of or and for and and is on of from the of the of the state of the and the of the AI is to from the or The of Executive Order the and as the there a in of making or the is the of such as think or in There are that and on AI speaking, the issue of algorithms. the National has National AI across the in with of their the issue of the National for and is for research, and for AI AI and and it can be to to the of algorithms. With to the with AI for information and the of has on facial recognition about and There are of from the scientific the and the The one is the National AI the use of systems in to are to the the of systems may from law to data and algorithms and it is that can To the that the on into with law of and is from the technology or it to in the In a has the to the data and algorithms to a in of the the of systems, are to input from the will the algorithm to the of the of a system to to the the decision to an decision making or decision making to There are no rules to use of algorithms. if the algorithm is to a rule, the of and The a as or a of an of or particular and to or law or policy or the or of any . algorithm that to with to or other be a does the decision making or the of Administrative are with of with the set in law, and other of by law, and on the set in law rules or other of are to as decision and The of on to of to AI and it is used in the of or With to there are of the one are with by that are often not to give an or if they the is of the used by the can be used by to the to and the law with the other the for are set in an the algorithm could be a rule, in it be to one of the In law, the to the and in a as well as the of in of the as and The turns more on the of and that be it does on the of the rules are with and are thought to on the circumstances of the AI the of the of by the and human of into the of the algorithm. the of is far there is the very that of in and other of will be from and will be by algorithms. is as a it the of and across the of that the of As a of law, there are no or law are in they to by or in their the if algorithms to more in such as a for they will be rules and they will have to the and of and of the use of systems limited to systems to and As of no such has in the law. With to has that there is a and of actual use and on for the of data or more a As one the to a but more in making may make it for or to the but may to and in the the law in Executive Order makes to the to the of and of law and other In the have the to the by the decision of the use and the of the does the to it a or the is to for a or the has a to the for the least one has that this of the data and algorithm used to the decision so that it is for the to an In the be by an of any the is to the of a that of the The have a to give of the data and the algorithm in the the law The law decision to an of the algorithm as of the and and the or the is not but the of and are and in all not of how an algorithm of that AI is the or of and by the and Office The machine learning to and and to for or are initial in the of a or application is The of are so that the of for any of or most does not an of it be that the of under and the on the circumstances of the The algorithm may be a that the of in the to give of the for the and the to the may have in by the algorithms may be used to in but they may and to In such the for is relatively the algorithm is from the decision and the for the to with the human and the and by the human of the the and the are to a to the by an There are no legal to or other of There the is the most from this is to make to about their the of a the their the of under the all of the that the of the in the used not and other In the of is the to on a data of the or are to the Act or the Act and the under there for of Executive Order that an of AI use this are to that their AI use are with the set in the They are to make their to the the of an algorithm under the to There has under the of Act on to the algorithms used by In the of to computer as it does to other of One of how used to to an algorithm is the of algorithms to a and to in to the be to by and of a of with and in the and in in the other they a of with limited data and a of with data the of and in of the for of the algorithm and for and the first it does not that the to the computer there are many to to and they often used by to to AI does to and with it in decision making or in the of the to the to As the of the law on from the and of and law. is the definition of in the of in decision There does not a legal definition of in the of algorithms. as to the of and AI but can be to on of the of in AI to and The in In with Executive Order and be a that in AI information about will and as well as and for the to information and in all of the to the and with legal legal on to, for and address or to are to the to the and and of frameworks or and the of other the of an system to an in or in There are of that to AI and are to or information technology that or information that is in an information that has to be to as information They do so a new of information that be or information and that that can be used for the of a . other the information on and and and The big data that is used to algorithms or not a on what it and it is are to a of new on the their use of Executive Order be to the Office of and in the Office of Management and and be by a a
Loading…
Imported from OpenAlex (CC0) — nothing here is generated.