How Should We Think About Ethics For AI?

It isn’t laws that keeps our societies functioning daily. It is ethics.

For that reason, I feel it is critical and time sensitive to explore and shape the ethical guardrails for some of the forms of AI usage.

Laws, by their very nature, are conservative in their reach and evolve slowly (for the most part).

Ethics is what guides the greater texture of behaviors, large and small, and shapes norms necessary for human systems to work. Without ethics, we would be in a constant state of unworkable litigation. This is why industries and institutions put in place ethics panels and give them authority to police behaviors.

Merriam-Webster Dictionary defines ethics as: “The discipline dealing with what is good and bad and with moral duty and obligation.” (note, there are multiple definitions dependent on context but I’ve chosen the broadest for discussion purposes.)

Ethics considers impact and fairness (to the extent that it can be defined), and asks of us to have care and regard for the effect of our behaviors on the working of institutions, social systems, and people’s lives.

I believe that the conversation around AI must also include a broad understanding of the ethical ramification of the tech, its impact on society, and the behaviors of those who will benefit most.

Presently, the discussion around AI seems to be focusing on the legality of its inner workings and its business plan. The laws, courts, and attorneys will decide those cases and we all have a lot riding on the outcome.

Ethics, however, is both an industry and public discourse. Any industry considering AI should be deeply engaged in discussions about the ethics of its use not just the legal ramifications. What those ethics are, and how they will be put in place, is yet unformed. That means we’re in dangerous territory because the rollout is well ahead of the ethics.

Getting in legal trouble should not be a company’s only concern about whether or not to do something. Especially if the impact of what you are about to do is so wide ranging and immediate.

Are your actions ethical and largely of good effect on the world? Have you involved the large number of people who will be heavily impacted in the discussion? Have you explored and identified the harm you could cause, intentionally or unintentionally? Have you formed the foundation for good ethics within your industry and are you transparent about those ethics?

From my vantage point, I do not see the AI industry engaging on these critical questions of ethics, nor do I see the client industries wholesale rising to get this done for themselves or requiring the AI industry to do so before adoption. I am sure there are a few companies beginning this thought process, but it is not a broad effort as of yet.

Given the rapidly growing impact on human, cultural, and commercial futures, I feel that the present approach and impact of the corporate rollout of mass commercialized AI is, in part, unethical. Ethics are a fuzzy conversation until they are broadly accepted and memorialized so here is my perspective.

?

1.?? With over 2,700 AI experts stating the there is a 5% chance that AI could cause the extinction of the human race, proceeding without extremely rigorous exploration and creation of robust limitations is unethical.

2.?? Willfully causing immediate and enormous profit-driven disruption through public release of AI with no equitable implementation plan, equivalent contribution to society, or public debate with the large numbers of those negatively impacted is unethical.?

3.?? Rapidly putting huge swaths of creatives out of jobs for a lesser product and concentration of corporate wealth is unethical (Please, let’s not have the disproven trickle-down conversation again.).

4.?? Vaguely hand waving about retraining of millions of people into tech related jobs, when they may not be suited for them, is unethical.

5.?? Damaging or eliminating the learning and career paths for young creatives in the university system is unethical (Creatives are just the beginning. Experts are projecting rapid job losses across many industries that will climb into the millions and more.).

6.?? Displacing truth, human voice, creation, and content from the larger society by flooding media with unlabeled AI content is unethical.

7.?? Inculcating data bias into the tidal wave of AI impact on human output is unethical.?

8.?? Building the foundation of large data sets on stolen materials then claiming ownership of the resulting model is unethical.

?

Net/net, we can be smarter about how we engage with AI on every level. ?Let’s elevate the ethics discussion and ask others to join us and engage on sorting this out now. Social push back works, and is far more effective incentivizing faster change. Waiting for the courts to do this work is slow and will not shape a largely beneficial rollout of this tech.

Cheers, Kevin

?