Written by: Matt Brown, Quantifi CTO
I’m involved in lots of conversations lately where the topic of machine learning or artificial intelligence comes up. These conversations include everything from self-driving (or flying) cars to advertising optimization to brain-computer interfaces.
When I started in software 20 years ago, the web was a nascent idea and no one even talked about data science. The idea that someone could predict when you might purchase something with enough confidence that they would move the product to the closest distribution center was science fiction. But today, the number of problems that can’t be solved by a machine is approaching zero. We haven’t solved them all, but we have practices and techniques that can tackle almost any problem imaginable.
Machines are getting better than humans at all kinds of problems including disease detection, crash avoidance, inventory management, package sorting. The list grows every day. I find the prospect of most these advancements very exciting especially in the areas of medicine and transportation.
In the midst of all these advancements, there is one thing that machine learning and artificial intelligence are not going to help with – the negative impact on people. As computers get better at driving vehicles, millions of truck drivers prepare for the eventual loss of jobs. As machines are able to better detect diseases and abnormality, radiologists question their career choices. Even software engineers need to start worrying. In the not too distant future, computers will write better software than we do. So what do we do about it?
This is a different kind of hard problem. Our new company is attempting to make marketing and advertising better through the use of machine learning. We could be part of the problem. But we want to be part of the solution. In fact, it’s important enough that one of our 6 core values is focused on the issue.
Create software that makes people and communities better
We’re creating tools in our platform that help marketers make better decisions by taking advantage of the machine learning we’ve built into our platform. Our hope is that the contributions of humans, combined with the power and scale of machines, will result in something better than we can imagine today. And that this will free people up to solve new problems.
But, our value statement has many layers. And the work we’re doing, and the possible effects, are complex. When you care about people and community, you have to also ask yourself many questions like:
- When a worker is displaced by software/computers, do you have a responsibility to them?
- Is a lack of low-skill jobs a good thing?
- Are efficiency and productivity the best metrics for determining usefulness?
- Just because you can doesn’t mean you should. Right?
And ultimately, what are the ramifications of what we’re building?
The Quantifi team is constantly considering and juggling these concerns as we strive for success and consider its costs.