In the face of calls for moratoria on AI research, it is\nclear that society wants to govern AI. \nIn the United States, we are governed by The Constitution. It lays out founding principles and\ninalienable rights, since amended 27 times, including the Bill of Rights. It defines the structure of our government\ninto 3 branches and how they are governed by the people. Thus, we govern ourselves, democratically. How shall we govern AI? (Presumably, we don\u2019t want to give AI the\nvote!)<\/p>\n\n\n\n
Recent advances in AI have shocked our social fabric. There is fear of economic upheaval, sudden\nsignificant unemployment, or deep fakes and toxic bots swarming social\nmedia. Fear is easy, and\nunderstandable. Thank you, Darwin, God,\net al. Flight commonly follows.<\/p>\n\n\n\n
The rational, educated, adult reaction to fear which is not\nimmediately threatening should be to think \u2013 without fear – before acting. Is the banning of AI by various social media\nsites, school districts, universities, etc. rational? <\/p>\n\n\n\n
We think \u201cyes\u201d in some cases, such as trying to avoid coding\nbots from overwhelming Stack Overflow with misinformation. AI is not smart enough to provide reliable\nanswers. And don\u2019t hold your breath,\nplease? The same is true for social\nmedia, where deep fakes and misinformation sewn by Russian bots can be\noverwhelming, but the more practical solution seems to be banning anonymity\nmore broadly. Once you know \u201cwho\u201d a user\nis, you can label or prohibit AI, but you can\u2019t tell which is which otherwise. To put that another way, unless you validate\nwho a social media user is, a ban is nothing but preening.<\/p>\n\n\n\n
We think \u201cno\u201d in education, generally speaking. In particular, trying to stop students from\nusing Chat GPT to help with homework or write assigned essays seems like a\nfool\u2019s errand. It\u2019s like banning\ncalculators. First, it\u2019s\nimpractical. The student will use the\ncalculator (or AI) at home and teachers will only find out when in-class\nbehavior differs. Second, the days of\nLuddites are past. Today, calculators do\nmuch more than arithmetic and their use is permitted even in standardized\ntesting, such as the SATs. Beyond\ncalculators, we don\u2019t have penmanship or typing classes anymore. The same will happen with AI in\neducation. It will fundamentally change\nwhat and how we teach and why it matters. \nIt may take as long as it did for calculators to be accepted, for\nprinting to dominate cursive, or for typewriters to disappear, but it\u2019s\ninevitable.<\/p>\n\n\n\n
Various proposals for a moratorium on AI development have\nbeen proposed. For the most part these\nare also fearful reactions. They seem\nmore rational, however, because they propose less action. They seek to freeze a moment in time to grasp\ntime to think. We think that\u2019s\ncounter-productive, but it\u2019s not unreasonable for society to make such a\ndecision. But it can\u2019t work for\nlong. And it will be costly.<\/p>\n\n\n\n
We think the rational path is to lay out a constitution\ngoverning AI. Just as the Founding\nFathers did not put society on hold while they debated and drafted the\nconstitution in Independence Hall, we prefer not to impose a moratorium on\nthose who are making a living advancing and applying AI towards staggering improvements\nin world-wide economics, education, healthcare, and more.<\/p>\n\n\n\n
Our corner of the AI community has adopted the notion of \u201cconstitutional AI\u201d promoted by Anthropic.\u00a0 This notion of constitutional AI requires a provider of AI to define the constitution by which it will abide.\u00a0 The dominant aspects of such constitutions today are that it will be helpful, harmless, and honest.\u00a0 However the AI provider defines the constitution, they represent to its users that it will substantially conform to its constitution and that the provider is committed to improving such conformance and ameliorating any non-conformance.<\/p>\n\n\n\n
That\u2019s all well and good, and we could go into the details of\nthe typical constitution at length, but it\u2019s not good enough for society. Society needs all such constitutions to be\ngoverned by a constitution that we, the people, accept. This shall be the constitution for AI.<\/p>\n\n\n\n
The debate by which this constitution shall be defined and\nratified belongs, as all matters of law, in the bodies of our legislatures,\ninfluenced as always by the public (and, of course, special interests). It should, as with all matters of\nlegislation, be out in the open and deliberate enough to allow for public\ndiscussion and protest before enactment, as may be appropriate.<\/p>\n\n\n\n
For decades, experts foretold superhuman AI as a decade\naway. Today, \u201cthe singularity\u201d, where AI\nbecomes completely general, capable of superhuman performance in all regards,\nis again, supposedly, a decade or so away. \nDon\u2019t believe the prognosticators. \nPast performance is indicative of future returns. <\/p>\n\n\n\n
Today, our legislators have little understanding of AI. They, like most of us, can imagine what AI is\nand will become capable of. None of us\ncan distinguish fantasy from reality when projecting AI forward.<\/p>\n\n\n\n
What matters is that quick legislation governing AI inside\nthe six-month moratorium some are calling for will be more fearful than\nthoughtful. If so, it will not allow\ntime for many of us, especially our legislators, and eventually regulators, to\nlearn the technical capabilities and limitations of the technology well enough\nto govern it wisely.<\/p>\n\n\n\n
How and whether other nations govern AI is not within our control, however.\u00a0 And is a perhaps the only rational argument needed to defeat calls for any moratorium.\u00a0 The reasons are somewhat obvious.\u00a0 In any case, given history, international consensus is ever evasive, implying that any moratorium will be either ineffective or perpetual.\u00a0 The former is catastrophic.\u00a0 The latter is impossible.<\/p>\n","protected":false},"excerpt":{"rendered":"
In the face of calls for moratoria on AI research, it is clear that society wants to govern AI. In the United States, we are governed by The Constitution. It lays out founding principles and inalienable rights, since amended 27 times, including the Bill of Rights. It defines the structure of our government into 3 … <\/p>\n