The 10 Commandments of AI
St. James Faith Lab (SJFL) has devoted a good deal of attention to neural networks, machine learning, and big data. Indeed, not so long ago in 2017, St. James welcomed a delegation from IBM to present its Watson machine-learning system to the congregation and some of the results they hope to be able to achieve with it. These are amazing and important new developments whose applications are already benefiting (or harming) us every day. In the longer term, however, their significance lies in the fact that they are stepping stones to achieving true artificial intelligence.
Artificial intelligence (AI) is poised to become one of (if not the) most significant scientific advancements in our history as a species. The ethical, moral, and tangible impact of the creation and eventual proliferation of AI is nothing short of staggering, and will completely upend our lives, and those of our children. As a theme explored by countless science fiction authors and creators, our media has been all but saturated by explorations of all the ways this could end up going sideways. Few stones have been left unturned: from subtle and nuanced explorations on the definition and meaning of life (memorable AIs include Star Trek’s Data and Ex Machina’s Ava) to disturbing explorations of ethical and moral questions that will need answers sooner or later (Blade Runner and its recent sequel, or Star Trek), all the way to apocalyptic worst-case scenarios (The Matrix, Superintelligence, Next).
We should not let ourselves dismiss these themes based on the popcorn entertainment value (or even lack of quality, in some cases) of some of the films and works that have explored this theme. The fact is, many of the world’s most recognizable names (Google/Alphabet, Microsoft, IBM, etc.) are heavily invested in making advances in AI, and while no one can say with certainty when the milestone of self-awareness will be reached, the Faith Lab believes it is right around the corner and will likely happen within our lifetimes.
Continued from the emailed newsletter
Given the magnitude of the moral implications of this milestone, the SJFL is concerned that Hollywood seems to be the most recognizable source of attempts to set and explore rules of ethics and morality that need to be fleshed out, and ideally set in place before a true AI is created. As things stand, we are essentially depending on the scientists and researchers themselves to set any such rule, with no consensus or external review. Yet, as most scientists would probably agree, morality generally falls outside the scope of scientific pursuit. Rather, this should be spearheaded by entities inherently devoted to morality: Christians, other religious communities, and non-religious ethical think tanks need to step up and devote thought and resources to these questions. We believe there is a need and a desire for such a guideline for those currently involved in the AI industry.
With these things in mind, the members of the SJFL have banded together to make our contribution to the discussion. We, much like the Church at large, are diverse, but share a common devotion to Scripture, and have particularly found inspiration in the 10 Commandments. As a starting point, we have assembled a list of the 10 Commandments of AI, with the ultimate goal of discussing and arriving at a clear set of guidelines for researchers and programmers dedicated to the development of Artificial Intelligence.
We list the 10 Commandments below, along with our interpretation of how it would apply to an artificial intelligence, and the resulting AI equivalent of each commandment.
1. You shall have no other gods before me.
One of the concerns of AI is that the ultimate goal is to unlock the secrets of the universe and be God, thereby rendering the human race “obsolete.” Even a (possibly inevitable) future in which AI develops capacities or abilities that vastly exceed what a human mind is capable of, no AI should ever treat a human as inferior, much as no person should either.
AI should never treat a human as anything other than sacred.
2. You shall not make idols.
AI has become the Holy Grail of this century casting aside religion. However, as a chiefly scientific advancement, AI should not be created to blur the line into spirituality and faith. We must never treat AI as superior to ourselves, much less as a deity to be obeyed or even worshipped.
AI shall not be worshipped.
3. You shall not take the name of the Lord your God in vain.
AI stands to be the first self-aware creation of the human race. This does not make its creators (companies or individuals) divine. We cannot think of ourselves as gods, nor should we create AI to make us feel like gods.
AI shall not be glorified.
4. Remember the Sabbath day to keep it holy.
AI should benefit the human race, and in turn we must commit to preserving our fellow humans and community values and to protecting the future that our children will live and share with any future artificial intelligence.
Remember, AI will affect our children.
5. Honor your father and mother.
A “good” AI should respect and promote fairness in all humanity despite its flaws, in all of its diversity of cultures and practices. Under no circumstances should it be made or used to perpetuate biases, increase or decrease social, racial, ethnic, economic, environmental, or sexual-orientation inequality, or promote one nation or people at the expense of another.
AI shall honor all creation.
6. You shall not murder.
AI, like any major scientific advancement, can conceivably be used, intentionally or not, for nefarious purposes. Any organization involved in its research and development should firmly commit to not apply it to developing more weapons or the means of creating or improving them or their manufacture.
AI shall not murder.
7. You shall not commit adultery.
AI should not be created or used with the intent of enabling or encouraging immoral or harmful activities on the part of humans, even legal ones. As an example, if an AI could conceivably be leveraged in a dating/matching app, this should not then be made to encourage people to hurt each other by cheating on their spouses (e.g.: an “Ashley Madison” AI). AI should be designed to prevent harm whenever it can and promote conditions for healing, collaboration and sustaining the greater good.
AI shall not enable fake or immoral acts.
8. You shall not steal.
AI should not be created or used with the intent of committing crimes, or to unjustly or illegally benefit itself or its creators at the expense of others. It should not be deployed to enhance cybercrime or commit theft.
AI shall not steal.
9. You shall not bear false witness against your neighbor.
One of the hoped-for benefits of an AI is that it could overcome the bias and prejudice that any human will inevitably struggle with, even subconsciously. We must not subvert or twist an AI to preserve biases, benefits, or privileges currently ingrained in society, or to promote falsehoods, “fake news,” or deepfake material such as imitating the speech or writing of another for purposes other than entertainment.
AI shall not bear false witness against its neighbor.
10. You shall not covet.
AI should not be limited to self-serving goals or to power any form of espionage, whether by corporations or by nation-states. Rather, it should foster openness and discovery, to the benefit of all.
AI shall benefit humankind.
What are your thoughts about the “10 Commandments of AI”? Let us know!
The Rev. Canon Cindy Evans Voorhees
St. James Faith Lab