A few weeks ago, I spoke at API Days Conference in New York. The title of my speech was “Solving The Maze of Technology Ethics.” Which, by the way, sounds somewhat mouthful.
But if you think about what this topic entails, what is this all about?
Is it about AI Ethics? Is it about Ethical AI? Is it about Responsible AI? Or, is it about Technology Ethics? While it is about all of the above, I think what this really about is being considerate.
Being considerate, because as a developer as a project manager as a system architect or anyone in this ecosystem, as a technologist, I think you have the power!
When I was a kid, I was always marveled by cars and how they function. But, I couldn’t quite figure out how cars actually move forward. So I asked my dad, and he explained to me how they operate. He also told me what the function of accelerator, brakes, the clutch is, and the steering wheel is. I still had one nagging question about the brakes. Why do we need brakes, wouldn’t they slow the car down? He continued and told me that the function of the brake is to slow down the car, but the purpose is to let it go faster.
And, that stuck in my brain forever. The brakes in the car do not slow it down; they allow it; they enable it to go faster! That’s the key message.
The brakes in the car do not slow it down; they enable it to go faster!
We have had many technology companies who followed move fast and break things model, and unfortunately, we have broken so many things. But then people always believe that having some policy, quality policy, corporate governance policy, or something similar would be helpful.
Unfortunately, merely having a policy is useless.
You see, the API ecosystem is adding to the exponential change very rapidly. Every single day a new API comes out, it affects hundreds of other APIs. We have to be careful about what we are developing and what its consequences are going to be.
If you want to develop a sustainable and a great product, it has to satisfy three criteria. The first criterion here is the solution must do the right thing, and we all know — every solution, every product starts with good intentions. We all want to do a good job, and I think that is not even a question here.
The next thing we need is doing it right. Fortunately, there are several avenues where you can learn how to develop APIs or other tools in the best possible manner — the best practices, the best technologies, etc.
The third part, though, which is the most critical one, and which is the most ignored one, is — covering all the risks. Most often, people and businesses flunk at it. They think covering the risk is all about project risk or business risk. However, it doesn’t limit there. We are talking about the risk that a solution over its entire period is going to pose. So, we are talking about unintended consequences here.
You may start with the best of the intentions, and you may want to do the right thing. You may have developed your APIs or your solutions in the best possible manner. But there are risks that you haven’t thought about, or considered and catered for. That is where those risks come into play. When you provide to those risks, when you address them, you mostly cover all the three points I mentioned above. Then you will have a five-star product in your hand.
Now the key here is, addressing those risks actively. Merely having a policy is not going to help. Again I say, it is going to be useless if you only have a policy. You need an actionable framework.
Merely having a policy is useless. You need an actionable framework!
When we saw several money laundering cases in the past, we always blamed banks. We thought that banks should have done a bit more. Banks should have placed specific measures to prevent money laundering.
I think, now, on similar lines, as a technology ecosystem, it is incumbent upon us to place similar measures to avoid these consequences. To prevent these risks, the right intention products may pose.
So, how do we do that? By including guidelines and actionable frameworks into process — design, development, deployment, and maintenance. The entire process needs to have these frameworks instilled in.
One of my favorite tools is pre-mortem analysis. I wrote about this in one of my latest books, Keeping Your AI Under Control (Apress, 2019). Let me explain a bit more about that.
Using a pre-mortem tool will help you in addressing unintended consequences. Pre-mortem is the hypothetical opposite of post-mortem. With the post-mortem, something has already gone wrong, and now we are looking at the thing asking what went wrong and what kind of damage it has done. How could we have prevented that damage, and how could we have stopped it from happening? What can be fixed? Etc. It’s mainly a post facto analysis as the damage is already done. It is now a damage control situation. Pre-mortem, on the other hand, is thinking about those failures or consequences and evaluating what could go wrong and why? It is done quite earlier in the design and development process.
I have written a separate piece on how to do pre-mortem analysis here: https://www.anandtamboli.com/blog?title=a-systematic-approach-to-risk-mitigation.
The point here is, you have to be quietly confident about your solution. And you get confident when you use these types of tools. By measuring the risk and by addressing it step by step, you become technically confident. You become convinced that you have covered all those risks.
Now you might be thinking, “Hey, if I go through these all hoops, it will slow me down significantly. It will stifle innovation because I can’t move fast!” So, let me repeat it, “The brakes in the car do not slow it down; they enable it to go faster!”
Just remember that maxim and be confident by measuring, by being cautious.
One thing we must remember that in the future, any organization that has high ethical capacity is going to have a competitive age. And, to develop that competitive edge, you must be confident. You must be developing sustainable and ethical solutions.
If you want to develop and build a sustainable future, a future, when you look back at you will be proud of. You must start implementing these practical frameworks in your entire process: design, development, and deployment processes.
Technology always brings us two things — it brings us a promise of prosperity and growth, and it also brings us consequences. It is the consequences that we must be worried about and concerned about. We have to be thoughtful and considerate about these consequences!
So, please be considerate, because you have the power!