Carlos Montenegro Murillo, Sr. Software Engineer at Mismo
Artificial Intelligence (AI) has certainly seen a revolution in the past couple of years thanks to significant increases in computing power and radical advancements in algorithmic performance. As such, it very quickly becomes a key component in almost every industry – providing immense value in any circumstance where large datasets are involved.
This has been incredibly exciting, and companies have been building out their AI capabilities as fast as they can, but the community has faced a lot of criticism in regard to the impact of these algorithms on society at large. It’s safe to say that we don’t have everything figured out just yet and the ethical implications of deploying AI algorithms at scale are immense.
Deploying AI ethically is a crucially important piece of the puzzle that needs to be taken seriously. If you are a technology start-up using AI, here are three of the key principles that you should be aware of when it comes to using this technology ethically:
- Critically assess your training data set. Regardless of what your machine learning system looks like, the quality of your data is the most important factor that will affect the end result. As such, it’s crucial that you evaluate the training data you are using to ensure it is fairly representative of the socio-economic conditions into which your solution is being deployed. You should do all you can to eliminate bias at this early stage so that it doesn’t get baked into the system.
- Set stringent algorithmic policies. While your engineers might like to be given free rein, it’s important that your leadership structures are able to set up and enforce AI policies that aim to protect the interests of your stakeholders. Your product development process should involve a sophisticated risk assessment component where you assess not just the technical risks but also the potential social and ethical implications of releasing your technology. Ethical checks and balances like these help to guide your company and ensure that what you’re building does not have negative externalities that will cause harm once in production.
- Hire the right engineers. The most important piece of this relates to hiring the right people. At the end of the day, those algorithms are in the hands of your engineers, and you want to make sure that they understand the relevant ethical considerations that affect your particular business. You should be looking to hire engineers who have empathy and an appreciation for AI ethics topics – who are willing to take these seriously. Seeing that this is a relatively new field, you may have to engage in some relevant training – but whatever you can do to train the engineers and data scientists to be ethical will help to avoid any potential errors down the road. You also want to be hiring a diverse set of engineers who can bring different backgrounds and points of view to the table because this also helps to avoid homogenous thinking that can lead to certain biases making their way into the product.
Those are just three principles to keep in mind to ensure that you’re deploying AI in a thoughtful and ethically aligned way. Here at Mismo, we facilitate outsourced engineering teams for clients all around the world and this is a key conversation in our organization. How do we empower our engineers to bring these topics up with clients and help to lead the industry in the right direction.
If you’d like to engage with us on this, or just to hear about how our remote engineers can take your business to the next level, get in touch today. We’d love to hear your story and see how we can help.