As you may already realize, AI has influenced your life. And it’s impact is only going to grow from here. Achieving a ubiquitous AI future could be life-changing. Think about never worrying about food nutrition and spoilage. Your milk never goes bad. Everything you eat is healthy. Or knowing exactly what clothes to pack on a trip, tailored to your occasion and weather when you arrive.
We’re not yet living in the dazzling future wherein a store immediately knows you walked in and starts suggesting custom-tailored products, but we may not be far off. This is why it’s crucial to understand — and break down — the roadblocks to AI adoption.
3 Roadblocks to AI Adoption
- Hardware and hardware compatibility is key and the technology isn't quite there yet.
- People have justified privacy concerns.
- As of right now, the necessary technology is very expensive.
The First Challenge: Hardware
Consider the smart fridge. Samsung introduced such an appliance in 2018, but it remains more of a novelty. According to Grandview Research, in 2019 “smart fridges” were most popular in North America, comprising 31 percent of the global market. Yet, these aren’t Star Trek-style devices with touchscreens, and instead are “smart” due to internal circuitry which allows for more efficiency and self-monitoring. The user may not even realize how “smart” the appliance is.
Major appliances are replaced on longer timeframes than your phone. Consumers replace large items as needed, so buying a new one just because it is slightly more efficient is less likely than getting a new phone because of a slightly better battery life.
Updating hardware also isn’t a trivial task. You can’t just add a Wi-Fi card to any fridge in hopes that it will broadcast a service record to the local repair center. Most of the electronics in our lives are not modular or meant to be extended much beyond their current design. This is a serious limitation to integrating internet of things (IoT) devices, as anyone with even just a couple of smart light bulbs will tell you. Hardware and hardware compatibility is key and we’re not quite there.
That said, a smart fridge — and other examples of smart hardware — are absolutely necessary to build the future of our AI-powered lives. It will take a moment for us all to get on the same page, with some of us buying into the future of AI-powered hardware sooner than others. The early birds will be crucial for leading the way to mass adoption, helping to work out the bugs and proving that such items not only work but add value to people’s lives.
The Second Challenge: Privacy
If you’re like me, the thought of AI knowing what I want to eat is kind of weird. We’re entering an era where our personal data will be more valuable than ever, and consumers are beginning to wake up to that fact. A report in 2019 indicates over 60 percent of respondents felt connected devices were “creepy,” which will likely slow adoption of such devices.
While all of this may sound daunting, there are some interesting innovations addressing the pain points. And you’re likely enjoying the benefits of this thinking without even realizing it. To understand, we have to go into a room filled with networking gear.
Most of us are familiar with server rooms thanks to TV shows and movies where we see some generic, but high-tech, “data center.” What most consumers don’t realize is that companies don’t just upgrade all their data center hardware at once. Just as you likely don’t buy a new router when you buy a new laptop, data center components are swapped out over time, here and there, and can wind up as a patchwork of vendors and services.
Some time ago, network administrators unified their management while allowing underlying systems to micro-manage the individual components. This requires special software that can amalgamate all the different requirements across all the different devices, controlling them as needed while obfuscating the details from managers.
As data centers are upgraded over time, more and more privacy is baked in. While we’re not quite in a place where we should trust our every move to an AI, we can expect most data centers to be privacy-centric in the next few years.
The Last Challenge: Cost
As you can imagine, nothing we’re talking about is cheap. The costs associated with current AI solutions are often prohibitive. This won’t always be the case, however.
Three decades ago, the computing power found in a smartwatch was as big and as expensive as a Toyota Corolla. Now you can get a $5 chip that is smarter than all the machines in all pre-2000 NASA space missions combined. The costs will fall.
We are already pushing AI to the edges — in more cost effective ways — by layering software on top of existing hardware instead of waiting on specialized AI-specific chips. We can add features to “dumb” machines by tapping into their networks and power grids.
Returning to our not-that-smart refrigerator, what if you replaced your electrical box with a smart electrical box that detected the fridge in your house based on its power usage? The smart power box would know the make and model and could make decisions about the fridge’s contents based on that (albeit imperfect) information. Add in a smart kitchen camera and maybe a counter-embedded scale and you’re adding sensors without adding much cost.
Ultimately, the best AI solutions will leapfrog all these roadblocks. They will get AI to the end consumer without relying on dedicated chips, which would require consumers swapping out new gear for old. After all, ubiquitous AI relies on it actually being where you need it. They will also be private from the start and as cheap as the air we breathe. Do any of us think about the Wi-Fi around us anymore? No. We use it and expect it. AI will soon be the same.