To give you some background, I have spent a couple of weeks this year visiting my friends who are raising young children between the ages 3 months -4 years old. While I was visiting them, I found that two of the households used a Google Home or Amazon Alexa to make their otherwise busy lives a bit manageable. The Home or Alexa would be asked to do anything from ‘Read the News’, ‘Check the weather outside’ to ‘Play a song/rhyme for the child’ , ‘Turn the lights off’ or ‘find the travel distance between Point A to Point B’. It was my first personal encounter with families using these home devices at such a high frequency. It almost felt like there was an invisible person present in the room at all times. I joked with one of my friends saying, ‘this should make it easy for you to explain the existence of God who knows it all, is present everywhere (via different devices) and is invisible at the same time’.
Why do I say you are raising a child, the AI child? Much akin to raising a human offspring which begins learning i.e. begins developing a brain in the mother’s womb through conversations/sounds/smells etc., the Google Homes/Google Search/Alexas of the world are deployed around us to catch all the random unstructured talk that occurs in human life. We are familiar with the term Big Data (simply, lots of data which can be both random and or specific) and Artificial Intelligence to some extent either through our work or news or plain interest in the subject. Artificial Intelligence is use of technology to make machines learn and become intelligent. For machines to learn, there is an abundance of data that’s needed. Machines aren’t smart to begin with, but we are teaching them how to learn and making them ‘intelligent’. Most machine learning talks about ‘neural’ networks like the ones humans have in their brains which grow as we learn leading to creation of synaptic networks (my psych and tech influence coming together after all😊). As the number of people using these devices increases and the frequency with which each person uses them increases, the more unstructured data is available for the development of intelligence. Our everyday conversations, commands, questions, requests are being absorbed by the AI brain, multiply this by the various languages people speak, tones, dialects, subjects (math questions, grammar, history questions etc.) and of course the millions of people using these tools. Why is this unstructured data? Simply said, it is random and may not always have a logical cause and effect relationship.
The question that arises here is do you know and think of this in terms of raising a ‘non-human’ being? Is this a wanted or unwanted child? Do we have a collective voice which holds any weight in choosing which of these technological advancements we truly want? I don’t think so. In my opinion we at least ought to know what we are getting ourselves into to the extent that we can as of today. Researchers in a lab may be aware of most ill effects of their inventions but they don’t always have the freedom to share it. Would you have wanted to know the potential harmful side effects of using DDT, RoundUp, Birth Control Pills, Genetically Modified Organisms besides what was declared on the label? I would want to believe Yes since that impacts our lives today and those of the future generations. Wi-Fi/Bluetooth and smartphones currently plague our lives, impact our sleep, hormonal balance and brain development. No one while developing it or selling it to us told us that the above technological advancements could be carcinogenic, cause health issues and also cause unintended harm to our planet. We were only told about the benefits.
The other question is, do you care to know? My short answer to this is, you may not care today but very soon you will need to when you/your children are forced to live with the consequences. We think that we are using the devices for our convenience but not many of us are consciously aware that we are also part of one of the biggest human experiments. Our life and life experiences are being fed to a machine so that it learns and grows. We spend our money to buy these devices and we also give free data to be used simply by using the device. We like to think that we are in control but often, we are not in control of how our data is being used. In most scientific experiments conducted, the ‘experimental’ group is aware that their observations/reactions/responses will be recorded as a part of the study. Are we as a consumer group made aware that our data is being used for non-human being machines to grow?
Am I against feeding data for machine learning? No. If I am not against it then where is this going? I am urging you to think about Ethics. Ethics of being a part of the experiment and ethics around the development of such artificial intelligence. The 3 main issues here are that of Informed Consent, Data Privacy, and Client/Consumer Welfare.
Informed Consent: From the experimental standpoint, if you are raising a child, you should have some rights, the least of it is to know that you indeed have a child and that you agree to raise it. This would be Informed Consent. I understand that Google owns all the data we have ‘shared’ with it’s applications but as an unaware participant in an experiment, I think someone ought to ask me before they decide to use the data from my Gmail Account/Facebook account to improve their AI. Am sure that some type of blanket or umbrella statement/disclaimer exists somewhere but that won’t cut it or shouldn’t cut it.
The second issue is about Privacy. Each of the applications nowadays come with a set of permissions you can deny/allow. Let’s be honest and admit that not many of us read through the privacy policies. We are shocked when we read articles that most online marketers are privy to data about consumers(us) which we may not always have explicitly agreed to share. Each application comes with some basic data it needs for you to sign up, for example, email address/google sign on/facebook sign on. Online marketers have access to data which connects these unrelated dots i.e. email address and name on one website with credit card, purchase data, address, age, income from another. How would it be if your doctor/lawyer/therapist said that he/she owns your data and has rights to share with anyone he/she pleases to simply because you visited for a consultation?
Even if you don’t care much for the above 2 I think you should care about Client/Consumer Welfare. Who gets to define what is welfare? The point here is that of instilling a sense of ‘Intentionality’ in the field of AI. An intention to create a non-human being which aids in human advancement and not harm us or be amenable to harmful uses. As of today, humans are constantly coding and improving AI and like most other human inventions, this is open to flaws/consequences-both intended and unintended in the future. Not all uses can be known when a code is being written but we do have enough experiential data to know that any innovation can be used to both help and harm our current existence. The differentiator is the intention of the creator and the user. There are some talks about wanting to develop super intelligence in AI beings. I have also read about wanting to upload intelligence from a living brain of a dead person into AI, so it could learn and develop faster. What are the safeguards in place if these AI beings are able to recreate their own versions in the future? Who should be held accountable for the current though process in development and the vastly open ground of unimagined uses in the future? If you and I are not explicitly agreeing to raise this child, would we assume responsibility for it’s consequences since silence at times is treated as consent? If we can have viruses/bugs that trump vaccines/medications created by humans, who is to say that this so called helpful AI technology won’t assert it’s will on us?
I have some ideas on how to proceed with instilling a more ethical approach in AI. This would of course need a lot more detailing/discussing when someone decides to implement any of these.
- Ethics and governance
- Like many other professions, Engineers should be asked to follow a code of Ethics – to do no harm and be a whistle blower when they can sense a potential misuse
- Some form of Board of Ethics utilizing members from all walks of life i.e. Doctors, Lawyers, Artists, Philosophers, Scientists, Teachers, Engineers who can
- debate the potential uses/misuses of a piece of code and disclose the same to the public
- define ‘Client Welfare’ for each Project.
- define a set of overarching principles which lay down basics of helpful/harmful uses.
These should be a rotating set of members who get randomly assigned to projects. This should help minimize some external influences from people who have monetary interests in the project.
- Once the framework of helpful and harmful uses has been established, public can vote on whether they would like to see the effort take life or not.
- Track the cognitive development of the AI body as you would track for a human child or develop something close to that.
- Until the AI body is considered to be a minor, the parent i.e. the company developing the AI should accept responsibility for any and every harm the AI body caused- known or unintended. In my opinion we should definitely begin this with self driving cars who are still developing their own sense of judgement.
- Once the AI body is deemed to be an Adult, implement a system of Shared responsibility between Users, Creators and the Creation (since it will have intelligence) . For example, if a user is attempting to use the digital technology for human trafficking then warn the user and allow him/her to rescind the process, if the user chooses to ignore the warning then share the user’s IP address/details with local law enforcement authorities. The creator i.e. the engineers/companies should be asked to code in a way that the AI body is able to determine if a use could cause harm to any person/group. The AI being should also be coached in/fed with data from philosophy, liberal arts, social sciences so that it can develop it’s own sense of helpful/harmful uses which is in alignment with overall human and planet welfare.
How to make this check and balance system ‘money and power proof’ will be a question that will need to be answered.
We know that AI is going to change our experience of being human as have other recent technological advancements. I shall leave you with a few thought clouds for you to ponder upon:
- Can we make an effort to be pro-active by involving diverse minds to foresee the risks and impacts?
- Can we think beyond the first order consequences and wonder about the psycho-social, spiritual and environmental impacts during the process of creation?
- Can we let common people have a say instead of simply forcing this new unforeseen age on everyone?
- The Universe in a Single Atom: The Convergence of Science and Spirituality he urges for the use of ethics in science and technology by Dalai Lama
- The Ethics of Artificial Intelligence by Nick Bostrom and Eliezer Yudkowsky; https://intelligence.org/files/EthicsofAI.pdf
- Ted Talk by Cathy O’Neil – The era of blind faith in big data must end; https://www.youtube.com/watch?v=_2u_eHHzRto
- Introduction to Deep Learning and Self Driving Cars- Lecture by Lex Fridman; https://www.youtube.com/watch?v=1L0TKZQcUtA
- How Google Marketers exploit your discomfort by Patrick Berlinquette; https://medium.com/s/story/make-orwell-fiction-again-part-2-micro-moments-9ba6e042a0c4
- Principles of Life and Work by Ray Dalio
- Counseling and Psychology theories – Transactional Analysis, Developmental stages
- Ethics in Healthcare
- My current experience of being human