AI Ethics: Is It Even Possible On A Global Scale?

We need to invest in cyber defences now!

The best definition of ethics I can find is …

the science of the ideal human character

As humans, we make mistakes in our lives, and which we then regret. Someone of good character is able to spot these things and try to improve themselves. Our strive in our lives must thus be for the idea of human character … of caring, of compassion, of patience, of respect, of tolerance, of integrity, and, generally, towards a world that removes greed, hatred, and spite.

But what about our AI children, who are growing fast? In just the last few years, they have grown up fast and are now better than us at many tasks. While they still struggle to move their skills from one domain into another, they are learning fast and learning all of the knowledge that has ever been created on the planet. but from us — their parents.

Can we make sure that our AI children have “the ideal character”? As good human parents, we try to bring up our kids to have a good human character. We show politeness and manners, and they are reflected in our offspring. But there are many others who bring up their kids to have bad characteristics. Of course, there are never any guarantees for this, and bad parents can bring up good offspring, and vice-versa.

So, why was Microsoft Tay taken off the Internet after just a few hours of life? Because she learnt to be a human from the worst human parents on the planet. And, remember that AI is learning not just from large-scale data sets that encapsulate every single human thought and item of knowledge but also from everything that we say and do in an online world.

AI Declarations on Ethics and Morals

The UN Universal Declaration of Human Rights defines a number of articles, including:

1. All human beings are born free and equal in dignity and rights. They are endowed with reason and conscience and should act towards one another in a spirit of brotherhood.

3. Everyone has the right to life, liberty and security of person.

But still, we have war, hatred and killing.

2. Everyone is entitled to all the rights and freedoms set forth in this Declaration, without distinction of any kind, such as race, colour, sex, language, religion, political or other opinion, national or social origin, property, birth or other status.

But still, we have horrible hate speech towards others.

And, so, any charter on AI will go the same way, and where the politicians and lawmakers will be able to state beautiful statements of intent, but AI will care little for that [here]:

And with the UN [here]:

Overall, the UN AI resolution defines a number of objectives:

• Safe, secure and trustworthy. “Towards AI systems that will also benefit sustainable development for all”
• Same rights, online and offline. ”To refrain from or cease the use of artificial intelligence systems that are impossible to operate in compliance with international human rights law”
• Closing the digital divide. “Do something about the varying levels of technological development between and within countries”
• Defining who governs AI. “We govern this technology rather than let it govern us”.

But, it is a technology that has never been in the hands of so many so quickly. In our society, we have many good people who believe in the principles of the UN Charter, but there are also many bad people in our world who are motivated by greed, hatred and aggression. Like it or not, it is those who wish to spy on us, take our money, and abuse us who are likely to be the winners in a game of AI ethics and morals.

In real life, we will be attacked like never before

In AI, we should have fundamental rights to privacy, and not to be spied upon for every second of our lives, but that’s not going to happen. For all the great that AI can do, its main application is likely to be on the battlefield, its usage by cybersecurity adversories, within the relentless strive to make us consume more, and in swaying our votes in an election. Brexit provided just a hint of how the Internet could be used to target every voter and provide them with fake news. This will be the tip of the iceberg, especially as the Internet has little respect for the borders that we have in place. Our whole notion of nation and regional states is likely to fall apart, but perhaps the EU will give us a role model of how the future world will look.

Politicians will put up the same old defensive mechanisms to protect “their patch” — that was a historic place that once saw battles to protect its existence. But, they will have little to defend themselves against the onslaught of bad AI — and with intelligence that will far outweigh the best legal minds on the planet.

What can we do?

But, what can we do? Well, just like with our human spirit, we need to strive for the good in our character and for this to be reflected in our AI. But, this will only be the AI entities that we can control. What about the rest? Well, we need to invest heavily in cybersecurity; otherwise we are heading for a world protected by George Orwell in 1984, or a world where we become slaves to the machine.

Conclusions

I repeat again … never in the history of humankind have access to such a powerful technology and for it to be in the hands of so many at the same time. Politicians and lawmakers have managed to constrain the rise of technology before. With AI, the have lost it, before they even had a chance to constrain it. We need investment in cyber defences, now! Otherwise, we will risk the future of our forthcoming generations. The attackers will not be dumb scripts and Kali Linux command lines but carefully crafted attacks that are created by an AI agent with (possibly) more intelligence than any person on the planet. It will know you better than you even know yourself.

Our history is one of greed and hatred; let’s, for once, do things properly for the planet and for those who come after us. AI is no real respecter of the physical borders we have placed between ourselves, but still, lawmakers see these false borders as the way they constrain anything. AI will learn that they have possibly caused more problems than they have done good in our world. We have harvested this planet for our own greed, and, if AI does some good, it will try and stop us from destroying ourselves, and build a better world. For this, we need to be ramping up our cybersecurity infrastructure to levels that have never been seen before.