Virtual Assistants need a soul

Virtual assistants, like Alexa and Siri, need a soul.

Not soul, like James Brown, although that couldn’t hurt, but a soul, like we all need. What is a soul? There are as many answers to that as there are people.

Defining what people’s souls have in common is important for maintaining our humanity and our control over the machines we create. The following are some probably lame generalizations.

Aspects of a person’s soul:

a) Integrity

In this context I interpret integrity to mean a striving to maintain a personal domain that is not interfered with by others. Integrity allows for interaction, but as little interference and as few side effects as possible.

b) History

This includes a persons memories, creations, resources, and other artifacts of their existence.

c) Individuality

This includes the rules and reasoning a person uses to judge or filter information. Some are innate but most are built up over time and are cumulative. This would include for example the ability to understand and speak different languages and dialects. This is a persons history incorporated into their senses. Since all of reality is so massive as to be impossible to experience and record, our individuality guides our experience and recollection.

d) Ethical code

This is also a highly individual part of a persons domain. These are the rules of the rules, and they vary not only across space and time but from person to person and even within people across time and context.

This list is far from complete.

Virtual assistants can understand speech and simple commands. So far as I know they don’t store history but that may be on the way. But as far as I know they don’t incorporate integrity, individuality or ethical codes.

How can we trust a virtual assistant that has no soul? I think the Solid project is a good beginning for the soul of personal assistants.

4 Likes

If viewed broadly, there are indeed as many answers as there are people because each individual soul is unique. But if viewed narrowly, every living, sentient organism has a soul, but no inanimate objects have one. Therefore, narrowly construed, a soul is binary; either one is alive, sentient, and thus has a soul, or one is a lifeless object. Of course, the question then becomes “what is life”? Is a virus alive? No matter how you answer that question, nobody will be heard to seriously argue that a virus is sentient. I think, therefore I am. A virus is a strand of genetic code covered by a protein sheath. A virus does not think, but unfortunately, it still is.

Cyberspace has become a reflection of ourselves and our psychologies. The Ego, SuperEgo, and Id (https://www.simplypsychology.org/psyche.html) now resemble the Open Web, Deep Web, and Dark Web. https://www.thesslstore.com/blog/what-is-the-dark-web. Social Media platforms like Facebook and Twitter manipulate human emotions for profit by abusing algorithms to create epistemic bubbles and eco-chambers, both of which have caused damages to humanity. https://philpapers.org/rec/NGUECA.

Tag42git defines a) Integrity, b) History, c) Individuality, and d) Ethical Code to be aspects of a person’s soul. A person should have integrity, but a government owes a higher standard of integrity than any one single person does. A person’s history is defined by their biology, their environment, and their contributions to humanity. Conversely, history will judge governments for leaving humanity and planet Earth in a worse condition than before they first came into existence.

Through the development of English common law, the rules and norms persons adopt to filter information have become the procedural and substantive laws enforced by governments. No matter what language or dialect one subjectively speaks, one should not, all else being equal, be treated differently by any one government. Cultural languages may differ, but mathematics is the one language which all living sentient beings can speak and understand. Ethical codes and norms are merely subjective rules which, according to English common law, become our by objective (judicial) rules then enforced by governmental judiciaries.

Virtual Assistants such as Alexa and Siri are created by third parties such as Amazon and Apple. Like the media and governments, Virtual Assistants must communicate effectively to their broadest audiences. If Virtual Assistants were designed to communicate only to people who have specialized skills, I suspect that the terminology which they listen for and react to would differ drastically, and thus so would their codes.

I also believe that the Solid project is a good platform upon which individualized Virtual Assistants may be designed to assist each person in all of their daily activities, to interact effectively with another person’s Virtual Assistant, to parse all facts of each person’s POD to minimize the potential for disputes between each person on both sides of that fence, and to the extent that disputes do arise, to resolve them automatically, electronically, and at no, or at least minimal, cost to each person.

So yes, humanity should have control over the machines that we create. But given the historical failures of all analog forms of governments, maybe its time that the machines we create have limited control over the forms of governments we imagine into existence. The one thing all human souls have in common is that each of us yearns to be free and to thrive. When Facebook, Twitter, and the like become so influential that they function like transnational governments, charting the future to a greater degree than any classical government had previously accomplished, it is time to rethink what social media and virtual assistants can do, it is time to rethink who should create them, for whom should they be created, and what they should be given the right to do for all of us; it is time for a Solid platform upon which we will design them.

Well said. I think we should be very careful about giving machines more control over our lives, and I include governments among machines. For example, there is a big difference between an electric car that you drive and a self driving electric car. The self driving one might be safer and more efficient, but how do we know? That’s I think where linked data comes in, to help us have more understanding and control over such things, not less. More than just linked data, Solid is a line that we won’t let machines cross.

Electric self driving cars avoid burning fossil fuels and leverage their connectivity to cyberspace in order to provide utility which the model T could never have predicted or accomplished in its day and age. Likewise, Benjamin Franklin, although undoubtedly an expert in electricity, could hardly have anticipated the existence of cyberspace when he contributed to the composition of the U.S. Constitution. For that same reason, it is difficult to predict how that future technology may be abused, such as hacking the internet to hijack that self-driving car in order to kill a person. This kind of crime could not have been previously predicted or accomplished with just a model T. Another example is, if a robot you build buys ecstasy on the internet, do you become criminally responsible? https://www.theguardian.com/world/2015/apr/22/swiss-police-release-robot-random-darknet-shopper-ecstasy-deep-web

Thank you for exposing the weak link in my analysis. Great care must indeed be taken not to give machines any discretionary control, but rather for us to have complete control over the machines which may have any influence upon our lives. To be certain, the dynamic norms which form the rules all governments enforce should be the only control the machines may have.

This is true for two reasons. 1) Tedious, yet perfunctory rule based tasks, are trivial to machines, yet we can do trivial things that are (currently) overly complex for machines to accomplish in this day and age 2) Those same machine “super-human” abilities can be leveraged to ensure a complete and transparent factual record should review become necessary. I will look further into incorporating linked data into my concept, but also how that data may lose value if linked in an unpredictable way. For example, if a contract was signed on a Tuesday (an objective fact all U.S. persons agree upon), the weak link in that fact should not prevent a judge who only knows the Chinese calendar from appropriately adjudicating a dispute which arises therefrom.