Eighty-five percent of jobs that will exist in 2030 have not yet been created. Artificial intelligence (AI) is one of the technologies influencing our future, including changing the way in which we access and analyze information in our everyday lives – business and personal. As the adoption of AI increases, what does that mean for non-engineers like me?
I recently sat down with Chris Stephenson, co-founder and CEO of Topos Labs, to discussed what the world would look like if AI could be democratized like PCs? Remember a time when only IT experts could set up our computers? Now, we can download apps and get ourselves set up – whether it’s a computer or a phone.
In this interview we focus on artificial intelligence, its growing popularity and why every business professional should be armed with the capability to use AI – not just engineers and technical experts.
Parna: Everyone talks about AI. Can you explain it simply?
Chris: There are many definitions of AI. But in the simplest terms AI technology simulates and automates human intelligence processes: learning, reasoning and self-improvement. Examples of AI include speech recognition, facial recognition, natural language comprehension and natural language generation.
Parna: What does democratizing AI mean?
Chris: Democratizing AI means making it easier to use and more broadly accessible to non-technical professionals, as opposed to a small group of highly trained engineers. At Topos, we often describe our platform Gracie as “AI for the rest of us.” Our goal is to empower the people on the front lines of business — analysts, researchers, consultants, domain experts — with this powerful technology, while reducing their reliance on IT, making them more autonomous, efficient and productive. And we practice what we preach. The team that trains Gracie mainly consists of Liberal Arts graduates with no formal programming or technology skills. Our core engineers never need to get involved in building our AI models. This is revolutionary stuff.
Parna: Why is it important to give more non-techie people access to AI?
Chris: Here’s an incredible stat: according to Forbes, 90% of the data that exists today was generated within the last two years. Data is growing exponentially, and AI is our only hope of organizing, understanding, prioritizing and protecting this data. So, beyond the benefits of empowering people, this really is a scale and skills issue. There are simply not enough machine learning engineers and programmers to go around. By simplifying this powerful technology for non-technical professionals we can scale the knowledge and optimize the productivity of our existing work force.
Parna: What three things need to be done to truly democratize AI?
Chris: The solutions must be simple, affordable and fast.
1. Ease of use is key. This requires an intuitive user interface. AI shouldn’t be an obstacle to the work, but an enabler of it.
2. Affordability is critical to allow broad access to companies large and small.
3. Speed is vital to keep up with the dynamic, ever-changing data and business needs. This means results in seconds and minutes, as people cannot afford to wait days or weeks to build and adjust AI models to solve their everyday needs. Those days are over.
Parna: Are there any risks to democratizing AI? If so, can they be mitigated?
Chris: As AI becomes more mainstream, there is a laundry list of risks and challenges to overcome. So of course, the concept of democratizing AI only heightens those risks.
Key risks I think about when making AI widely accessible center around the improper or unethical use of the technology. Fortunately, there are ways to mitigate these risks:
1. Accuracy: The solution must provide the ability to quality-check, optimize and improve the models to ensure the results are credible and useful for the business.
2. Explainability and Transparency: The solution must not be a “black box” but include methods to understand how the predictions were made, and to allow the user to identify and fix potential data bias or pollution.
3. Education: It is important to provide helpful documentation, guidance and support for the solution to ensure the system is used properly to avoid confusion or unintended consequences.
One final thought: While AI is increasingly allowing us to do more with less and tackle previously impossible challenges, we need to remember it’s just a tool, meant to augment and support us in our work and our decisions. Keeping a proper perspective is critical to maintaining a healthy human/machine balance.