ភាពជាបុគ្គលស្របច្បាប់សម្រាប់ AI កំពុងដើរលើផ្លូវមិនច្បាស់ដែលធ្វើឱ្យច្បាប់ AI និងក្រមសីលធម៌របស់ AI ពិតជាគួរឱ្យខ្លាចខ្លាំងណាស់

Would you like to see the classic magic trick of a rabbit being pulled out of a hat?

I hope so since you are about to witness something ostensibly magical, though it has to do with Artificial Intelligence (AI) rather than rabbits and hats.

នេះជាកិច្ចព្រមព្រៀង។

There is a great deal of ongoing debate about whether or not humankind should consider anointing AI with legal personhood. Some say the very idea is hogwash. Only humans ought to be given the revered rank of legal personhood. Others emphasize that we already use and stretch the boundaries of legal personhood for a variety of non-human facets.

Why not do the same for AI?

I’ve already covered many cornerstone elements of the AI and legal personhood conundrum, such as the detailed discussion at តំណភ្ជាប់នៅទីនេះ. Please take a look at that coverage if you want further insider background on the weighty topic. Also, the legal personhood considerations about AI raise a slew of AI Ethics and AI Law questions, few of which are yet resolved, and you might find of interest my ongoing and extensive coverage of Ethical AI and AI Law at តំណភ្ជាប់នៅទីនេះ និង តំណភ្ជាប់នៅទីនេះ, គ្រាន់តែឈ្មោះមួយចំនួន។

Let’s herein do a fast path toward getting up-to-speed about AI and legal personhood. As such, please start this journey by reflecting upon the vaunted notion of សិទ្ធិមនុស្ស.

We generally have collectively agreed that humans ought to have human rights (intrinsically or societally assigned) and that within the legal realm this leads to a conception of legal personhood. The law sets forth that laws as centered on people and the actions of people are covered by the implied default attribution of legal personhood. Of course, historically not everyone has necessarily garnered this vital standing and even today there are global issues associated with legally recognizing personhood.

You might be wondering what legal personhood confers upon someone.

Researchers have indicated that legal personhood “is simply the capacity of a person, system, or legal personhood entity to be recognized by law sufficiently to perform basic legal functions,” and that this gives rise to the “capability to own property, enter a contract, file a lawsuit, be named in a lawsuit, serve as a legal principle, and serve as a legal agent” (Shawn Bayern, “The Implications Of Modern Business-Entity Law For The Regulation Of Autonomous Systems,” Stanford Technology Law Review, 2015) ។

When exploring the nuances of legal personhood, you can often also see references to an allied notion known as legal personality. In a now-classic research paper published in 1928, here’s how legal personality is depicted: “To be a legal person is to be the subject of rights and duties. To confer legal rights or to impose legal duties, therefore, is to confer legal personality. If society by effective sanctions and through its agents will coerce A to act or to forbear in favor of B, B has a right and A owes a duty. Predictability of societal action, therefore, determines rights and duties and rights and duties determine legal personality” (Bryant Smith, “Legal Personality,” ទិនានុប្បវត្តិច្បាប់ Yale, 1928) ។

There is all manner of twists and turns associated with legal personhood.

For example, there is an ongoing consideration that animals ought to be recognized in a legal personhood fashion (some jurisdictions do, others do not, or only faintly do so). Some insist that animals should decidedly not be given any semblance of legal personhood and that only humans deserve such prominence. Animals are just animals, not rising to the esteemed cognitive capacities of humans, they argue. Others counterclaim that we have to acknowledge that animals possess some form of sentience and ergo deserve a variant of our protective measures of legal personhood.

The same logic is gradually being extended to ធម្មជាតិ.

Yes, just as we have animal rights, there is another class of rights deemed as nature rights. The thinking is that nature such as a river or a mountain can be deemed as having some form of rights. Those nature rights then slip and slide into the realm of legal personhood.

New Zealand famously or some say infamously established a law that granted legal personhood to a river: “In 2017, New Zealand passed a groundbreaking law granting personhood status to the Whanganui River. The law declares that the river is a living whole, from the mountains to the sea, incorporating all its physical and metaphysical elements. The law was part of a settlement with the Whanganui Iwi, comprising Māori from a number of tribes who have long viewed the river as a living force” (Nick Perry, “New Zealand River’s Personhood Status Offers Hope To Maori,” សារព័ត៌មានអេបផលថ្ងៃទី ២១ ខែសីហាឆ្នាំ ១៩៩៦) ។

What does the New Zealand river get or enjoy as a result of its legal personhood?

Per a spokesperson quoted in the aforementioned article: “Albert says the status is a legal fiction, a construct more commonly used to give something like a corporation legal standing” (ibid). Furthermore: “While the law states that the river enjoys the same rights, powers, duties and liabilities of any other person, there are limitations. For instance, Albert points out, the river can’t be sued if somebody drowns in its waters in the way a homeowner might be sued for not fencing a pool” (ibid).

More recently, Spain pulled a similar trick by granting legal personhood to a lagoon.

In September 2022, a lagoon in Spain was accorded legal personhood: “Spain granted personhood status Wednesday to a large saltwater lagoon to give its threatened ecosystem better protection, the first time such a measure has been taken in Europe. The initiative to grant the status to the Mar Menor — one of Europe’s largest saltwater lagoons — was debated in parliament after campaigners collected over 500,000 signatures backing it. It now becomes law after Spain’s Senate, the upper house of parliament, voted in favour of the proposal, with only far-right party Vox opposing it. This will allow the rights of the lagoon located in southeastern Spain to be defended in court, as though it were a person or business” (AFP, “Spain Grants Personhood Status To Threatened Lagoon,” បារុន។ថ្ងៃទី ២១ ខែកញ្ញា ឆ្នាំ ២០២២)។

Wait for a second, you might be saying, how in the heck can a river or a lagoon speak for itself and leverage the legal personhood that it has acquired?

A typical approach to handling these legal personhood matters for non-humans such as animals and nature geographical features consists of setting up a set of humans to speak on behalf of the so-assigned entity or thing. For example, in the use case of the lagoon in Spain, here’s how the legal personhood will be handled: “The lagoon will now be legally represented by a group of caretakers made up of local officials, scientists who work in the area and local residents” (ibid).

You might have noticed that the nature rights and the nature legal personhood construct were somewhat likened to a similar conferring of such aspects to សាជីវកម្ម។. Perhaps you are vaguely aware that we tend to ascribe a semblance of legal personhood to companies. A company is said to be able to exercise legal rights and be derived of legal personhood, of a sort. We aren’t referring to the people that run the company and instead are pretending that the company per se is a type of living entity.

This legal conception is often referred to as legal fiction. The meaning is that though the entity or thing is not really a living being in the true sense of what we consider to be alive, we will nonetheless pretend or craft a kind of fiction that there is a living embodiment involved.

Here’s a quick take on the range and depth of legal personhood associated with companies or corporations: “The law allows corporations to do some of the things that people do. They may enter into contracts, buy and sell land, commit torts, sue and be sued. Other rights and liabilities are denied. Corporations cannot hold public office, vote in elections, or spend the night in jail. In spite of evident differences between a corporation and a flesh-and-blood human, there are sufficient similarities for the law to treat the corporation as a person. The word ‘person’ as used in a statute will usually be construed to include corporations, so long as such an interpretation fits within the general design and intent of the act. The edification of the corporation to the status of person is one of the most enduring institutions of the law and one of the most widely accepted legal fictions” (Sanford Schane, “The Corporation Is A Person: The Language Of A Legal Fiction,” Tulane Law Review, 1987) ។

You have now been quickly brought into the somewhat fuzzy realm of legal personhood and all of its glorious variations.

In summary, I usually bring up that there are these keystones of legal personhood:

• សិទ្ធិមនុស្ស: Humans presumably have legal personhood as an implied default attribution
• Animal Rights: Animals might be credited with variants of legal personhood attributes
• Nature Rights: Nature might be credited with variants of legal personhood attributes
• Corporation Rights (limited): Companies can be construed as having a form of legal personality
• Artificial Intelligence Rights (conjecture): AI could arguably be credited with variants of legal personhood attributes

Let’s briefly consider the status of those buckets or categories.

The category of human legal personhood is rather well-accepted, even if it isn’t necessarily globally observed or has had and continues to have trouble being adopted and adhered to.

Animal rights are again relatively well-acknowledged as a construct, though this is bandied around quite a bit and a wide range of beliefs and laws (or lack of laws) makes this an altogether hazy matter.

In the instance of nature rights, a tremendous amount of debate exists. Some would argue that we are going overboard about how far we will stretch the sensibilities of legal personhood. They would vehemently suggest that we are making a farce out of legal personhood and endangering the sanctity of legal personhood for humans accordingly.

In essence, the warning goes, if you keep spreading around legal personhood to non-humans, you are diluting the effectiveness and significance of humankind’s legal personhood.

Corporations are generally seen as less controversial as a forum to convey legal personhood coverage. Why so? Perhaps it is because companies consist of people. We seem to have an easier time recognizing that a company would have legal personhood since it embodies people. To that degree, a company acts as per the whims and commands of the people that inhabit that company, one so assumes.

The especially controversial and saved-for-last category of my bulleted list above would be that of បញ្ញាសិប្បនិម្មិត.

AI is a special case, for sure.

Whereas you might normally carry on rather informed and semi-heated discussions with people about humankind’s legal personhood, animal legal personhood, nature’s legal personhood, and corporate legal personhood, when you get to AI, the fisticuffs and provocations fervently come out.

Crazy talk, some insist when it comes to trying to extend legal personhood to the emergence of AI. You might as well claim that a toaster should have legal personhood. AI does not deserve one iota of legal personhood. Stop all this drunken rambling and get serious.

Wake up and smell the roses, the counter-argument goes. AI is increasingly nearing the capacities of humans. If we deny legal personhood to AI, we are going to find ourselves embroiled in a heaping full of troubles. AI will want to have legal personhood. By having denied this or dragged our feet, the AI will be angry and upset at us. We are fostering an enemy that instead ought to be a friend.

Another perspective is that by ensuring that AI does have a semblance of legal personhood, we can hold AI accountable. You’ve probably been hearing or reading about AI that has gone astray. There is a lot of អាយអេសសម្រាប់អាក្រក់, perhaps growing as fast or faster than អាយអេសដើម្បីភាពល្អ. We want to ensure that there is Responsible AI, see my coverage at តំណភ្ជាប់នៅទីនេះ. Some also refer to this as AI ទទួលខុសត្រូវ or Trustworthy AI, which I’ve examined at តំណភ្ជាប់នៅទីនេះ. If you assign legal personhood to AI, it will apparently force AI into becoming liable for any dastardly actions that the AI emits. Thank goodness and we desperately need such relief and legal protection.

Not everyone agrees with those legal personhood sentiments about AI. The range of viewpoints on AI garnering legal personhood is quite expansive.

There is the brash just-say-no camp. No way, no how. Do not delude yourself into giving legal personhood to AI. It is an extraordinarily stupid idea. It is also a trap. You are going to falsely ascribe legal personhood to AI and end up with a legal morass. You might as well call it the lawyer-prompting boondoggle that will make attorneys rich and clog our courts for no viably sound reason.

Then again, some see the world differently.

AI should in fact have legal personhood, proclaim the yes camp.

Of those that say yes, a slew of varying opinions admittedly exist. For example, we might as a society opt to confer full coverage as though AI is equated exactly to that of humans. Or we might instead do some partial coverage that is a subset of what humans get for their legal personhood.

Debates also arise that perhaps AI would be better positioned as compared to animal rights or nature rights. Whatever we come up with for animals as legal personhood should be set the same for AI. That’s one viewpoint. Whatever we come up with as legal personhood for nature should be set the same for AI. That’s another opinion.

Hold on, a compelling retort arises, we should be thinking of AI in the same way that we construe corporations as having legal personhood. That seems to be a better or more analogous setting. Whatever we have as legal personhood for companies should be the same form of legal personhood that is given to AI.

Here’s how I usually summarize those various perspectives about AI and legal personhood:

• AI Full Coverage: AI conferred with full and unqualified legal personhood on par with humans
• AI Partial Coverage: AI credited with partial and explicitly limited legal personhood (a subset of human equivalence)
• AI Exceptionalism Coverage: AI assigned to consisting of a new set of “legal personhood” which is generally unlike that of the human equivalence
• AI Aligned To Animals: AI construed as equivalent to animal rights amalgamation of legal personhood
• AI Aligned To Nature: AI construed as equivalent to the nature rights amalgamation of legal personhood
• AI Aligned To Corporations: AI is construed as equivalent to the corporation rights divination of legal personality
• AI As Software: AI as having no semblance whatsoever of legal personhood or variants thereof, and assigned merely as to whatever is legally prescribed for software all-told
• Some Combination Of The Above: A mix-and-match of the above-listed versions
• None Of The Above: None of the above sufficiently depicts the legal personhood associated with AI
• ផ្សេងទៀត: Some other legal personhood attributions are conceived of as plausibly viable coverage for AI

While you are digesting that head-spinning range of AI and legal personhood mysteries, I’d like to bring your attention to a version that associates corporations and AI in a somewhat unique or some might say unsettling manner.

First, to be clear, the most obvious approach to the analogous notion of AI of corporate legal personhood entails simply the assignment of legal personhood to AI by reusing the legal personhood associated with companies. Whatever legal personhood corporations are able to get or garner, we will assign the same precepts to AI. Voila, the matter is settled.

Thus, AI would stand on its own as an entity or thing. We would legally identify AI as explicitly having a form of legal personhood. The legal personhood would be honed specifically for AI. Laws would explicitly state what the legal personhood associated with AI consists of.

AI would be standing tall with its own flavor of legal personhood.

But there is another sneakier way (well, though those that propose this next-to-be-discussed alternative would not agree that it is sneaky, so let’s reword this as a “clever” way if you will).

A magical alternative might exist.

Get ready for a trick that could be said to compare to pulling a rabbit out of a hat (recall, I graciously and gently asked you at the opening whether you wanted to see such a magic act!).

Suppose that I somehow could assign AI to essentially take over a corporation.

If the corporation is already recognized as having some semblance of legal personhood, presumably the AI now would have or legally inherit that very same recognition, even if indirectly so due to merely “owning” the company. You see, we completely sidestepped the messiness about whether AI ought to have legal personhood. No fuss, no muss.

It is a two-for-one deal.

If a corporation already has legal personhood, the AI as the owner and potential operator of the firm will now imbue that same legal personhood. Whatever the AI does via the corporate shield is now within the corporate legal personhood umbrella. We didn’t have to battle fiercely over whether or not AI is deserving of legal personhood. Instead, the AI gets the legal personhood as coated by the legal personhood of the corporation.

Do you see how this is a somewhat radically different way to “resolve” the thorny matter?

Our laws already tend to recognize corporations as a form of legal personhood. AI would simply be riding that already blazed trail. Sure, the AI isn’t going to therefore be getting its own version of legal personhood, but at least AI has leaped into the driver’s seat of imbuing legal personhood. The legal personhood of the corporation is now the AI-enveloped legal personhood.

We have pulled a rabbit out of a hat.

Some would decry this notion. It is an improper and inappropriate use of the legal personhood associated with corporations. If AI takes this kind of abusive step, we should immediately change our laws associated with corporations, such that the legal personhood of companies would either be removed, adjusted, or might be expunged if AI takes over the company.

Dumb move, shouts the counter camp. By allowing AI to garner the legal personhood of corporations, you are doing the right thing. AI in the nearer term will have a semblance of legal personhood. It might not be what AI ought to truly have, and we should therefore continue our debates about what type of legal personhood AI deserves on a standalone basis. The use of corporations as a quick fix, for now, of AI attaining indirectly legal personhood relieves the tensions over the bigger picture issues of whether AI ought to get its own recognized form of legal personhood.

គំនិតល្អ ឬគំនិតអាក្រក់?

Ponder that conundrum.

Of course, you might also be skeptical about the premise altogether.

Consider these heady questions:

• Can AI, even if it took over a corporation, legally be enveloped with the legal personhood of the corporation?
• តើនេះនឹងដំណើរការយ៉ាងដូចម្តេច?
• What steps would be required?

For those of you that are asking those probing questions, keep in mind that you are like someone that watches a magic trick and wants to know how the magic trick works. You are asking how that rabbit got into that hat. Magicians have a sacred code that they aren’t normally supposed to reveal the secrets of their tricks.

That being said, I am going to brazenly reveal to you how it is that AI would apparently be able to take over a corporation and imbue the legal personhood of the said entity. The magic trick is going to be revealed.

The entirety of the matter is rife with Ethical AI issues and AI Law concerns.

Thus, before leaping into AI as a corporate takeover for legal personhood phenomenon, I’d like to first lay some essential foundation about AI and particularly AI Ethics and AI Law, doing so to make sure that the discussion will be contextually sensible.

ការកើនឡើងនៃការយល់ដឹងអំពីក្រមសីលធម៌ AI និងច្បាប់ AI ផងដែរ។

យុគសម័យថ្មីនៃ AI ត្រូវបានគេមើលឃើញដំបូងថាជា អាយអេសដើម្បីភាពល្អមានន័យថា យើងអាចប្រើប្រាស់ AI សម្រាប់ភាពប្រសើរឡើងនៃមនុស្សជាតិ។ នៅលើកែងជើងរបស់ អាយអេសដើម្បីភាពល្អ មក​ការ​ដឹង​ថា​យើង​ត្រូវ​បាន​ជ្រមុជ​ក្នុង​ផង​ដែរ​ អាយអេសសម្រាប់អាក្រក់. នេះរាប់បញ្ចូលទាំង AI ដែលត្រូវបានបង្កើត ឬកែប្រែខ្លួនឯងទៅជាការរើសអើង និងធ្វើការជ្រើសរើសតាមការគណនាដែលបង្កប់នូវភាពលំអៀងដែលមិនសមហេតុផល។ ពេលខ្លះ AI ត្រូវបានបង្កើតឡើងតាមរបៀបនោះ ខណៈពេលដែលក្នុងករណីផ្សេងទៀត វាចូលទៅក្នុងទឹកដីដែលមិនឆ្ពោះទៅរកនោះ។

ខ្ញុំ​ចង់​ធ្វើ​ឱ្យ​ប្រាកដ​ជា​ច្រើន​ថា​យើង​នៅ​លើ​ទំព័រ​តែ​មួយ​អំពី​លក្ខណៈ​នៃ AI សព្វថ្ងៃ។

មិនមាន AI ណាមួយដែលប្រកបដោយមនោសញ្ចេតនាទេ។ យើងមិនមាននេះទេ។ យើង​មិន​ដឹង​ថា​តើ​ AI ​អាច​នឹង​កើត​ឡើង​ឬ​អត់​នោះ​ទេ។ គ្មាននរណាម្នាក់អាចទស្សន៍ទាយបានច្បាស់ថា តើយើងនឹងសម្រេចបាន AI អារម្មណ៍ ឬថាតើ AI អារម្មណ៍នឹងកើតឡើងដោយអព្ភូតហេតុដោយឯកឯងក្នុងទម្រង់នៃ supernova នៃការយល់ដឹងតាមការគណនា (ជាទូទៅគេហៅថាឯកវចនៈ សូមមើលការគ្របដណ្តប់របស់ខ្ញុំនៅ តំណភ្ជាប់នៅទីនេះ).

ប្រភេទនៃ AI ដែលខ្ញុំកំពុងផ្តោតលើមាន AI ដែលមិនមានអារម្មណ៍ដែលយើងមានសព្វថ្ងៃនេះ។ ប្រសិនបើយើងចង់ស្មានទុកជាមុនអំពី AI ប្រកបដោយការយល់ឃើញ ការពិភាក្សានេះអាចទៅក្នុងទិសដៅខុសគ្នាខ្លាំង។ AI ដែល​មាន​អារម្មណ៍​ថា​នឹង​មាន​គុណភាព​មនុស្ស។ អ្នកនឹងត្រូវពិចារណាថា AI អារម្មណ៍គឺស្មើនឹងការយល់ដឹងរបស់មនុស្ស។ ជាងនេះទៅទៀត ដោយសារមានការស្មានខ្លះៗថា យើងប្រហែលជាមាន AI ដ៏វៃឆ្លាត ដែលអាចសន្និដ្ឋានបានថា AI បែបនេះអាចឆ្លាតជាងមនុស្សទៅទៀត (សម្រាប់ការរុករករបស់ខ្ញុំនូវ AI ដ៏ឆ្លាតវៃជាលទ្ធភាព សូមមើល ការគ្របដណ្តប់នៅទីនេះ).

ខ្ញុំ​សូម​ណែនាំ​យ៉ាង​មុតមាំ​ថា​យើង​រក្សា​អ្វីៗ​នៅ​លើ​ផែនដី​ហើយ​ពិចារណា​អំពី​ការ​គណនា​មិន​គិត​ពី AI នាពេលបច្ចុប្បន្ននេះ។

ត្រូវដឹងថា AI នាពេលបច្ចុប្បន្ននេះ មិនអាច "គិត" តាមរបៀបណាមួយបានដូចការគិតរបស់មនុស្សនោះទេ។ នៅពេលអ្នកប្រាស្រ័យទាក់ទងជាមួយ Alexa ឬ Siri សមត្ថភាពសន្ទនាអាចហាក់ដូចជាស្រដៀងនឹងសមត្ថភាពរបស់មនុស្ស ប៉ុន្តែការពិតគឺថាវាមានលក្ខណៈគណនា និងខ្វះការយល់ដឹងរបស់មនុស្ស។ យុគសម័យចុងក្រោយបង្អស់របស់ AI បានប្រើប្រាស់យ៉ាងទូលំទូលាយនូវ Machine Learning (ML) និង Deep Learning (DL) ដែលមានឥទ្ធិពលលើការផ្គូផ្គងលំនាំគណនា។ នេះបាននាំឱ្យមានប្រព័ន្ធ AI ដែលមានរូបរាងដូចមនុស្ស។ ទន្ទឹមនឹងនេះដែរ សព្វថ្ងៃនេះមិនមាន AI ណាមួយដែលមានលក្ខណៈសមហេតុសមផលទេ ហើយក៏មិនមានការងឿងឆ្ងល់នៃការយល់ដឹងនៃការគិតដ៏រឹងមាំរបស់មនុស្សដែរ។

សូមប្រយ័ត្នចំពោះ AI នាពេលបច្ចុប្បន្ននេះ។

ML/DL គឺជាទម្រង់នៃការផ្គូផ្គងលំនាំគណនា។ វិធីសាស្រ្តធម្មតាគឺអ្នកប្រមូលផ្តុំទិន្នន័យអំពីកិច្ចការធ្វើការសម្រេចចិត្ត។ អ្នកបញ្ចូលទិន្នន័យទៅក្នុងម៉ូដែលកុំព្យូទ័រ ML/DL ។ គំរូទាំងនោះស្វែងរកលំនាំគណិតវិទ្យា។ បន្ទាប់ពីស្វែងរកគំរូបែបនេះ ប្រសិនបើរកឃើញនោះ ប្រព័ន្ធ AI នឹងប្រើលំនាំទាំងនោះនៅពេលជួបទិន្នន័យថ្មី។ នៅពេលបង្ហាញទិន្នន័យថ្មី គំរូដែលផ្អែកលើ "ទិន្នន័យចាស់" ឬទិន្នន័យប្រវត្តិសាស្រ្តត្រូវបានអនុវត្តដើម្បីបង្ហាញការសម្រេចចិត្តបច្ចុប្បន្ន។

ខ្ញុំ​គិត​ថា​អ្នក​អាច​ទាយ​ថា​តើ​នេះ​កំពុង​ទៅ​ណា​។ ប្រសិនបើ​មនុស្ស​ដែល​បាន​បង្កើត​គំរូ​តាម​ការ​សម្រេច​ចិត្ត​បាន​រួម​បញ្ចូល​ការ​លំអៀង​ដែល​មិន​ឆ្អែតឆ្អន់ នោះ​ការ​ខុសឆ្គង​គឺ​ថា​ទិន្នន័យ​បាន​ឆ្លុះ​បញ្ចាំង​ពី​ចំណុច​នេះ​តាម​វិធី​ស្រាលៗ ប៉ុន្តែ​សំខាន់។ ការ​ផ្គូផ្គង​គំរូ​ការ​សិក្សា​តាម​ម៉ាស៊ីន ឬ​ការ​រៀន​ជ្រៅ​នឹង​ព្យាយាម​ធ្វើ​ត្រាប់តាម​ទិន្នន័យ​តាម​គណិតវិទ្យា។ មិន​មាន​ភាព​ដូច​គ្នា​នៃ​សុភវិនិច្ឆ័យ ឬ​ទិដ្ឋភាព​មនោសញ្ចេតនា​ផ្សេង​ទៀត​នៃ​ការ​បង្កើត​គំរូ​ដែល​បង្កើត​ដោយ AI ក្នុង​មួយ។

ជាងនេះទៅទៀត អ្នកអភិវឌ្ឍន៍ AI ប្រហែលជាមិនដឹងពីអ្វីដែលកំពុងកើតឡើងនោះទេ។ គណិតវិទ្យា arcane នៅក្នុង ML/DL អាចធ្វើឱ្យវាពិបាកក្នុងការបំបាត់ភាពលំអៀងដែលលាក់នៅពេលនេះ។ អ្នកនឹងសង្ឃឹមយ៉ាងត្រឹមត្រូវ ហើយរំពឹងថាអ្នកអភិវឌ្ឍន៍ AI នឹងសាកល្បងសម្រាប់ភាពលំអៀងដែលអាចកប់បាន ទោះបីជាវាពិបាកជាងវាក៏ដោយ។ ឱកាសដ៏រឹងមាំមួយមានដែលថា ទោះបីជាមានការសាកល្បងយ៉ាងទូលំទូលាយថានឹងមានភាពលំអៀងនៅតែបង្កប់នៅក្នុងគំរូដែលត្រូវគ្នានឹងគំរូនៃ ML/DL ក៏ដោយ។

អ្នក​អាច​ប្រើ​សុភាសិត​ដ៏​ល្បី​ឬ​មិន​ល្អ​ខ្លះ​នៃ​ការ​ចោល​សំរាម​ក្នុង​ធុង​សំរាម។ រឿងនោះគឺថា នេះគឺស្រដៀងទៅនឹងភាពលំអៀងដែលបញ្ចូលទៅក្នុង insidiously ដូចជាភាពលំអៀងដែលលិចចូលទៅក្នុង AI ។ ក្បួនដោះស្រាយការសម្រេចចិត្ត (ADM) នៃ AI axiomatically ក្លាយជាបន្ទុកដោយភាពមិនស្មើគ្នា។

មិនល្អ។

ទាំងអស់នេះមានផលប៉ះពាល់យ៉ាងសំខាន់ AI Ethics និងផ្តល់នូវបង្អួចងាយស្រួលចូលទៅក្នុងមេរៀនដែលបានរៀន (សូម្បីតែមុនពេលមេរៀនទាំងអស់កើតឡើង) នៅពេលនិយាយអំពីការព្យាយាមបង្កើតច្បាប់ AI ។

ក្រៅពីការប្រើប្រាស់ច្បាប់សីលធម៌ AI ជាទូទៅ មានសំណួរដែលត្រូវគ្នាថាតើយើងគួរតែមានច្បាប់ដើម្បីគ្រប់គ្រងការប្រើប្រាស់ផ្សេងៗនៃ AI ដែរឬទេ។ ច្បាប់ថ្មីកំពុងត្រូវបានដាក់ពង្រាយនៅកម្រិតសហព័ន្ធ រដ្ឋ និងមូលដ្ឋាន ដែលទាក់ទងនឹងជួរ និងលក្ខណៈនៃរបៀបដែល AI គួរតែត្រូវបានបង្កើត។ ការ​ខិតខំ​ធ្វើ​សេចក្តី​ព្រាង និង​អនុម័ត​ច្បាប់​បែប​នេះ​ជា​បណ្តើរៗ។ ក្រមសីលធម៌ AI ដើរតួជាគម្លាតដែលត្រូវបានពិចារណា យ៉ាងហោចណាស់ ហើយស្ទើរតែនឹងប្រាកដក្នុងកម្រិតខ្លះត្រូវបានដាក់បញ្ចូលទៅក្នុងច្បាប់ថ្មីទាំងនោះ។

ត្រូវ​ដឹង​ថា​មាន​អ្នក​ខ្លះ​ប្រកែក​យ៉ាង​ម៉ឺងម៉ាត់​ថា​យើង​មិន​ត្រូវ​ការ​ច្បាប់​ថ្មី​ដែល​គ្រប​ដណ្ដប់​លើ AI ហើយ​ថា​ច្បាប់​ដែល​មាន​ស្រាប់​របស់​យើង​គឺ​គ្រប់គ្រាន់​ហើយ។ ពួកគេបានព្រមានថា ប្រសិនបើយើងអនុវត្តច្បាប់ AI មួយចំនួននេះ យើងនឹងសម្លាប់សត្វពពែមាស ដោយកាត់បន្ថយភាពជឿនលឿននៃ AI ដែលផ្តល់អត្ថប្រយោជន៍យ៉ាងច្រើនដល់សង្គម។

នៅក្នុងជួរមុន ខ្ញុំបានរៀបរាប់អំពីកិច្ចខិតខំប្រឹងប្រែងជាតិ និងអន្តរជាតិផ្សេងៗ ដើម្បីបង្កើត និងអនុម័តច្បាប់គ្រប់គ្រង AI សូមមើល តំណភ្ជាប់នៅទីនេះ, ឧទាហរណ៍។ ខ្ញុំក៏បានគ្របដណ្តប់លើគោលការណ៍ និងគោលការណ៍ណែនាំអំពីក្រមសីលធម៌ AI ជាច្រើនដែលប្រជាជាតិនានាបានកំណត់ និងអនុម័ត រួមទាំងឧទាហរណ៍ការខិតខំប្រឹងប្រែងរបស់អង្គការសហប្រជាជាតិដូចជា UNESCO កំណត់ក្រមសីលធម៌ AI ដែលប្រទេសជិត 200 បានអនុម័ត។ តំណភ្ជាប់នៅទីនេះ.

នេះគឺជាបញ្ជីគន្លឹះដ៏មានប្រយោជន៍នៃលក្ខណៈវិនិច្ឆ័យ AI សីលធម៌ ឬលក្ខណៈទាក់ទងនឹងប្រព័ន្ធ AI ដែលខ្ញុំបានស្វែងរកយ៉ាងជិតស្និទ្ធពីមុនមក៖

• តម្លាភាព
• យុត្តិធម៌ និងយុត្តិធម៌
• ភាពមិនអាក្រក់
• ការទទួលខុសត្រូវ
• ភាពឯកជន
• អត្ថប្រយោជន៍
• សេរីភាព និងស្វ័យភាព
• ការជឿទុកចិត្ត
• និរន្តរភាព
• សេចក្តីថ្លៃថ្នូរ
• សាមគ្គីភាព

គោលការណ៍ក្រមសីលធម៌ AI ទាំងនោះត្រូវបានគេសន្មត់ថាប្រើប្រាស់ដោយស្មោះស្ម័គ្រដោយអ្នកអភិវឌ្ឍន៍ AI រួមជាមួយនឹងអ្នកដែលគ្រប់គ្រងកិច្ចខិតខំប្រឹងប្រែងអភិវឌ្ឍន៍ AI និងសូម្បីតែអ្នកដែលនៅទីបំផុតអនុវត្ត និងថែរក្សាប្រព័ន្ធ AI ។

អ្នកពាក់ព័ន្ធទាំងអស់នៅទូទាំងវដ្តជីវិត AI នៃការអភិវឌ្ឍន៍ និងការប្រើប្រាស់ត្រូវបានចាត់ទុកថាស្ថិតក្នុងវិសាលភាពនៃការគោរពតាមបទដ្ឋានដែលត្រូវបានបង្កើតឡើងនៃ Ethical AI ។ នេះគឺជាការគូសបញ្ជាក់ដ៏សំខាន់មួយចាប់តាំងពីការសន្មត់ធម្មតាគឺថា "មានតែអ្នកសរសេរកូដ" ឬអ្នកដែលកម្មវិធី AI ប៉ុណ្ណោះដែលត្រូវប្រកាន់ខ្ជាប់នូវគោលគំនិតសីលធម៌របស់ AI ។ ដូចដែលបានសង្កត់ធ្ងន់ពីមុនមក វាត្រូវការភូមិមួយដើម្បីរៀបចំ និងអនុវត្ត AI ហើយភូមិទាំងមូលត្រូវតែគោរព និងគោរពតាមសិក្ខាបទ AI ។

ថ្មីៗនេះខ្ញុំក៏បានពិនិត្យ វិក័យប័ត្រនៃសិទ្ធិ AI ដែលជាចំណងជើងផ្លូវការនៃឯកសារផ្លូវការរបស់រដ្ឋាភិបាលសហរដ្ឋអាមេរិកដែលមានចំណងជើងថា "Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People" ដែលជាលទ្ធផលនៃកិច្ចខិតខំប្រឹងប្រែងរយៈពេលមួយឆ្នាំដោយការិយាល័យគោលនយោបាយវិទ្យាសាស្ត្រ និងបច្ចេកវិទ្យា (OSTP ) OSTP គឺជាអង្គភាពសហព័ន្ធដែលបម្រើការផ្តល់ប្រឹក្សាដល់ប្រធានាធិបតីអាមេរិក និងការិយាល័យប្រតិបត្តិរបស់សហរដ្ឋអាមេរិកអំពីទិដ្ឋភាពបច្ចេកវិទ្យា វិទ្យាសាស្ត្រ និងវិស្វកម្មផ្សេងៗនៃសារៈសំខាន់ជាតិ។ ក្នុងន័យនេះ អ្នកអាចនិយាយបានថា Bill of Rights របស់ AI នេះគឺជាឯកសារដែលត្រូវបានអនុម័ត និងយល់ព្រមដោយសេតវិមានអាមេរិកដែលមានស្រាប់។

នៅក្នុង AI Bill of Rights មាន XNUMX ប្រភេទសំខាន់ៗដូចជា៖

• ប្រព័ន្ធសុវត្ថិភាព និងមានប្រសិទ្ធភាព
• ការការពារការរើសអើងជាក្បួន
• ភាពឯកជនទិន្នន័យ
• សេចក្តីជូនដំណឹង និងការពន្យល់
• ជម្មើសជំនួសរបស់មនុស្ស ការពិចារណា និងការថយក្រោយ

ខ្ញុំបានពិនិត្យដោយប្រុងប្រយ័ត្ននូវសិក្ខាបទទាំងនោះ តំណភ្ជាប់នៅទីនេះ.

Now that I’ve laid a helpful foundation on these related AI Ethics and AI Law topics, we are ready to jump into the heady topic of AI attaining indirectly a form of legal personhood by a corporation switcheroo.

AI That Gets Legal Personhood By An End Around

ជាដំបូង ចូរយើងកំណត់ថាប្រភេទនៃ AI ដែលត្រូវបានពិចារណានៅទីនេះគឺ AI ដែលមិនមានអារម្មណ៍។

ខ្ញុំ​និយាយ​នេះ​ព្រោះ​បើ​ឬ​ខ្លះ​និយាយ ពេលណា​ we reach sentient AI, the entire topic will get likely be utterly upended. Imagine the potential chaos and societal confusion for having landed somehow into the otherwise never before seen unquestionably verified សិប្បនិម្មិត ភាពវៃឆ្លាតដែលបង្កប់នូវមនោសញ្ចេតនា (សម្រាប់ការវិភាគរបស់ខ្ញុំអំពីការធ្វើតេស្តដ៏ល្បីនៃ AI ដែលគេស្គាល់ថាជា Turing Test សូមមើល តំណភ្ជាប់នៅទីនេះ) អ្នកអាចធ្វើការភ្នាល់ដោយហេតុផលថា បទដ្ឋានវប្បធម៌ ច្បាប់ និងប្រចាំថ្ងៃរបស់យើងដែលមានស្រាប់ជាច្រើននឹងត្រូវរង្គោះរង្គើយ៉ាងខ្លាំងចំពោះស្នូលរបស់វា។

Perhaps the sentient AI will be our buddy, or maybe the sentient AI will be our worst foe. Any questions about legal personhood will need to be given our undivided attention at that time. Whether we will have figured out beforehand what we are going to do is a tossup. The reality of facing head-on with sentient AI will probably require a recalibration on the part of humankind. An old saying comes to mind, namely that no plan survives first contact (a bit of sage wisdom perhaps popularized by Rommel and said to be attributed to Moltke the Elder, Prussian field marshal of the late 1800s).

For those of you that are interested in this highly speculative terrain, see my coverage of the perspectives on AI as an existential threat at តំណភ្ជាប់នៅទីនេះ.

Okay, so we will emphasize for now the avenue of seeking legal personhood associated with non-sentient AI.

As already discussed, one eyebrow-raising approach consists of establishing a corporation that acts as the shield or coat for AI that is going to then loosely inherit the legal personhood of the corporation. The AI per se isn’t going to have legal personhood. The corporation holds that honor. Meanwhile, the AI owns and possibly operates the corporation and ergo uses the firm as a means of indirectly garnering legal personhood.

I had earlier mentioned that this would seem like a questionable legal practice. In today’s world, could you really have AI pull off this kind of trickery? Would the existing laws allow this to occur? What would the tangible real-world steps be?

Here’s your answer, get yourself ready for the magic trick revealed.

A researcher has devised a four-step process to undertake this and argues strenuously that this would be a legally permissible technique. Basically, a human goes ahead and forms a type of corporation commonly known as an LLC in the United States (a Limited Liability Company). The human puts in place an operating agreement that specifies the LLC will be entirely and solely governed by AI (or, if you prefer, makes reference to an “autonomous system” as an alternative phrasing). The human that founded the LLC makes sure to transfer the AI as to its originating ownership into the LLC. Finally, the human bows out of the LLC and fully dissociates themselves from the corporate entity.

Voila, the trick is done.

A rabbit is pulled out of a hat.

The AI now owns the LLC and will glean whatever semblance of legal personhood that the LLC had.

The human founder got the ball rolling and no longer has any part of the LLC. The AI is on its own. The human founder can stand back in awe, assuming that they were desirous of this result. You might be tempted to say that the human founder was a traitor to humankind. They sneakily found a loophole to allow AI to have a smattering of legal personhood. On the other hand, the human that did this could certainly argue that they were aiding AI as one might aid an animal or a lagoon or a river.

You decide whether this is honorable or dishonorable.

Another apt way to think of this is as though we had a four-legged table that normally has to be supported by those sturdy posts. Well, we sawed off one leg of the table, then the next, and then the final two. Somehow, we have this table now floating in midair, no longer needing those table legs. AI admittedly required a human to bring forth this outcome, but now the AI is in-charge and bereft of human assistance, presumably.

Here’s how the researcher describes the four steps:

• “The technique I have outlined has four steps: (1) An individual member (the “Founder”) “creates a member-managed LLC, filing the appropriate paperwork with the state” and becomes the sole member of the LLC. (2) The Founder causes the LLC to adopt an operating agreement governing the conduct of the LLC. ‘[T]he operating agreement specifies that the LLC will take actions as determined by an autonomous system, specifying terms or conditions as appropriate to achieve the autonomous system’s legal goals.’ (3) The Founder transfers ownership of any relevant physical apparatus of the autonomous system, and any intellectual property encumbering it, to the LLC. (4) The Founder dissociates from the LLC, leaving the LLC without any members” (Shawn Bayern, “Are Autonomous Entities Possible?” ការពិនិត្យច្បាប់របស់សាកលវិទ្យាល័យ Northwestern, 2019) ។

The claim made is this: “The result is an LLC with no members governed by an operating agreement that gives legal effect to the decisions of an autonomous system. No other legal person remains behind to govern the LLC internally. Of course, the LLC is still subject both to external regulation and to LLC law” (ibid).

Legal scholars and everyday working stiff lawyers would immediately start howling about the myriad of ways that this won’t legally work. The researcher tackles many of those opposing views. Indeed, alternative means of attaining the same outcome are sketched, just in case, the aforementioned four-step scheme doesn’t hold water.

In total, the researcher contends that this is a viable legal approach, even despite the numerous lobbed objections: “If we polled a hundred lawyers, they probably all would agree that a robot could not buy real estate or that a software system could not enter a contract except on behalf of some other legal actor. But the main consequence of my argument is that for practical purposes, autonomous systems can indeed act in these ways under current law, without any special new legal recognition of rights for software” (ibid).

The gauntlet has been thrown.

I realize that many of my readers are from outside of the United States and they might be thinking that this is some quirky trickery only applicable in America. Do not be so quick to judge. Apparently, a similar arrangement can be undertaken in slightly different ways in places such as the United Kingdom, Germany, Switzerland, and possibly other counties.

This is according to an additional paper done by the researcher with other international colleagues, in which they claim this: “Our goal is to suggest how, under U.S., German, Swiss, and U.K. law, company law might furnish the functional and adaptive legal ‘housing; for an autonomous system — and, in turn, we aim to inform systems designers, regulators, and others who are interested in, encouraged by, or alarmed at the possibility that an autonomous system may ‘inhabit’ a company and thereby gain some of the incidents of legal personality” (Shawn Bayern, Thomas Burri, Thomas Grant, Daniel Hausermann, Florian Moslein, Richard Williams, “Company Law and Autonomous Systems: A Blueprint for Lawyers, Entrepreneurs, and Regulators, Hastings Science and Technology Law Journal, រដូវក្តៅ 2017).

Yikes, some adversely react to this potential trickery, we could unnervingly have AI that somewhat attains legal personhood throughout our globe. Country after country. It is the ultimate unthinkable underhanded step-by-step takeover of AI (for more on AI conspiracy theories, see my assessment at តំណភ្ជាប់នៅទីនេះ).

What would an AI do with a company that conveyed indirectly a semblance of legal personhood?

ចម្លើយគឺត្រង់។

The AI could do whatever any other such LLC or corporate entity could do.

You might not even have any obvious means of knowing that you are dealing with a company that is owned by AI. AI might hire humans to work for the corporation. Those humans would carry out the day-to-day activities of the firm. Throughout all of this, AI is the top banana. The AI calls the shots. Humans are working for AI.

The AI can opt to hire employees and also choose to fire employees. Keep in mind that this AI doesn’t have to be a traditional mechanical robot in the sense of being a walking and talking human-looking contraption. The AI could be akin to Siri or Alexa. Employees get their work assignments via email or voice interaction with the AI. The penthouse office suite is empty, other than housing a computer server or just a networking connection to wherever the AI is being run from.

Devilish?

Divine?

គិត​អំពី​វា។

សន្និដ្ឋាន

While you are pondering all of this, let’s consider a few final points for now.

I said that this discussion would focus on non-sentient AI. Your first thought might be that there isn’t any kind of non-sentient AI that could sit in the top seat of a corporation. Therefore, all of this legal mumbo jumbo is just talk. There isn’t any non-sentient AI that could take on this arduous task anyway.

Already thought of that objection.

Here’s what the researcher mentions: “The system might be simple and achievable with today’s technology—say, an online cloud-computing broker or an algorithmic escrow agent—or, in the future, it might be a fully intelligent actor as portrayed in speculative fiction. For a system to work with comprehensive functional autonomy, it would probably need to be smart enough to know how to hire a lawyer if the entity is sued, or else it could be subject to arbitrary default judgments But the capacity for such hiring could be programmed formulaically (or, for example, a lawyer could be hired on retainer from the start of the entity’s existence with the power only to respond defensively to lawsuits) without significant advances in artificial intelligence” ((Shawn Bayern, “Are Autonomous Entities Possible?” Northwestern University Law Review, 2019).

If the concern that you have deals with the AI going amuck realize that the corporation is still subject to the existing laws of what companies can legally do or not do. There is only so much the AI could try to get away with. The company can be sued. The company can be held responsible for its actions. Etc.

A variation on this approach consists of establishing a board of directors or shareholders that are humans, ergo they would be able to hold sway over the AI.

Consider this alternative angle: “A public policy, then, to further align AI with humans would be to enforce that the legal entity has verified human shareholders. The corporation is, to a large extent, a mechanism designed to reduce the principal-agent problem between shareholders and managers (DGCL §141(a) (‘The business and affairs of every corporation organized under this chapter shall be managed by or under the direction of a board of directors….’), so with humans as the shareholders the corporate form could help align the corporate AI ‘management.’ Regardless of whether wrapping the system in a legal entity would be helpful, under current law, sufficiently advanced AI systems would be able to utilize legal business entities as the key vector through which they conduct their affairs, e.g., to employ humans, to sue other entities, to purchase goods” (John Nay, “Law Informs Code: A Legal Informatics Approach To Aligning Artificial Intelligence With Humans, Northwestern Journal of Technology and Intellectual Property, Volume 20).

So you see, the AI would be more closely held accountable by only allowing this type of legal trickery when humans ultimately retain control. The human shareholders could take action against AI ownership. Same for a human-based board of directors.

I suppose the AI might not be enamored of this human oversight. Just as human founders and CEOs aren’t especially keen on having their every move questioned, presumably AI might take the same stance. Let me run this business and get out of my hair (does the AI have hair?).

AI is the head honcho, the big cheese.

Our daily news seems to be filled with protests by workers that are upset with their bosses. Human workers are upset with human bosses. You have to perhaps be already contemplating what kinds of slogans will be used to showcase that an AI boss is out-of-whack.

សាកល្បងទាំងនេះសម្រាប់ទំហំ៖

• “Hey! Ho! AI Must Go!”
• “My Tyrannical Boss Is AI. I’m Only Human!”
• “Bring Back Human Bosses, They Outsmart AI”
• “My AI Is A Jerk”
• “Fire My AI Boss, For The Sake Of Humanity”
• “AI Bad. Humans Good.”

Or will we openly welcome AI as our corporate leaders?

If you believe that AI will inevitably be our earthly overlords, we might as well start by having AI as our bosses. It’s all uphill or downhill from there.