AI Doesn’t Work in the Real World
Just another case of not living up to the hype…
Well actually it does, and very well. It is just invisible to us and not part of what we experience as we use it. AI or more specifically narrow ai has been all around us for a 10 to 20 years already. It is embedded in some of our favourite experiences like Google, Netflix, and Amazon. These companies all did two things differently though. [Some as early as the 90’s!!] First they never led or even mentioned “how” these things worked. Two they slowly introduced customers to the experiences and built upon them based on customer engagement.
Artificial Intelligence, artificial general intelligence, full ai, narrow ai, machine learning, advanced analytics, predictive analytics, and all kinds of other impenetrable terms are being thrown around in popular media today. The reality is “ai” has been here for a long time and is behind the core products of some of the largest market cap companies in the world.
What’s strange is you never hear a kitchen knife company talking about how the steel is blended and then hardened during the knife forging process. The focus is where it should be, on the end product. That is the knife, not the method used to make the knife. Why all of the attention on AI which is the method, not the end product? Perhaps it is the novelty of it or the extreme prominence tech has taken in our personal and business lives. I would love to hear your thoughts on the this.
This is something I encounter almost every day and a week never passes without some of the friction from the current AI hype or the friction where people take a stance of people vs. analytics, often taking a combative stance.
Let’s pause the existential threat conversation for a bit, please
This isn’t the first time we’ve been here either. AI was being talked about between the science community and the media in the 70’s and 80’s when computing power started to enable more advanced applications. It even spawned a bunch of movies at the time. Everyone remembers the original Terminator movie, right? Skill testing question, do you know who HAL is?
My work experience has centered around product, design, and analytics. This has given me a lot of perspectives, both good and bad. More importantly, it makes the journey of continuing to improve the situation very important to me. I believe the human-centered design approach is what is needed here. It really does all start with empathic understanding. My view is that crossing the AI adoption chasm (now that the “cat is out of the bag”) is about engaging people in ways far beyond just the rational. We need to make these interactions engaging for people and valuable for people.
Where do we start though? AI today is utilitarian. Much more akin to a cognitive version of computing than any sort of self-aware entity. It shouldn’t matter in the context of product building in my opinion but there is a big difference between AI (or narrow-ai) and AGI or artificial general intelligence. With the practical tools, we already have, with AI we can now program machines with models for problem-solving and not have to just program solutions directly. It’s like teaching a child by asking them good questions versus giving them the answer. Did anyone else go through Montessori?
There are three guiding principles I am using in building my current product.
- It is human to want to feel part of the process. AI needs to earn the right to play.
- AI isn’t the end product, it is a method for crafting end products.
- The real world is messy and it goes well beyond data messiness.
So let’s get started…
It is human to want to feel part of the process. AI needs to earn the right to play.
Products that try to “automate out” human interaction have and will continue to have a hard time. The automation should be a partnership and allow for interaction from both sides. Products that adapt and become more useful through time are very valuable but we need to design a way for us to also play a conscious role in the product.
AI isn’t the end product, it is a method for crafting end products.
We can create amazing human and emotional experiences which are powered by AI. Much of the work before has looked at AI as a replacement analytics technique only. It is actually a decision-making method which can be used in the background to create new experiences.
Although AI is exceptional at complex analytical tasks like voice to text translation or translation between language there are so many other possibilities. For me, the one I can’t stop thinking about is all the new products we can design with the adaptive capabilities given to us by AI techniques. The possibilities for continuous customization and personalization, driven by AI, are huge. They are already behind Google search as well as Netflix and Amazon product recommendations.
The real world is messy and it goes well beyond data messiness
Efforts to date have focused on handling messy data. This is not enough. Models need to be useful but often imperfect as well.
The big thing here is better is what is needed not perfect. The question is does this improve from where we are today? If yes then use the capability now. This may be too analytical but we are never perfect or always right. How we perceive the world is much more about ourselves than it is about the world. AI should offer an advantage when it does use it.
One great way to get started is by imagining how you would design a product from the ground up with AI versus trying to fit AI into an existing product paradigm.
What next?
If you are a product manager, designer or data scientist (still not sure I like that term) this is your call to action! The onus is on us to make this work. It is NOT in the customer’s list of jobs to be done. AI is one of our tools in crafting products that meet their actual jobs to be done. The reward is in meeting that goal, not in how we got there. We can and should use the right to for the job, the simplest or most complex required, and then make that invisible to the customer. Make it “just work”.
Not one of us has all the answers but we all need to quiet ourselves, to hear the messages from our customers. We also need to look at their engagement and ways to improve it, not hand over the product reigns to them. Each of us has something unique to offer, the effort is in finding ways to engage people while showing them the value. Showing is much more powerful than telling. Let them experience it and you will both gain something.
This requires a growth mentality, of being committed to the mission of always getting better. Allowing the gains made and potential future gains to guide you rather than success or failure today. You can’t and won’t win at everything. The biggest rewards never come from a known path and the known path will never catch the pace of the those walking the unknown path forwards who are humble enough and confident enough to constantly adapt. Remember every product, every method is about the impact it creates in the world, not the product or method itself.
- It is human to want to feel part of the process. AI needs to earn the right to play.
- AI isn’t the end product, it is a method for crafting end products.
- The real world is messy and it goes well beyond data messiness.
Make it simple for your customer and achieve their goal!
“That’s been one of my mantras — focus and simplicity. Simple can be harder than complex; you have to work hard to get your thinking clean to make it simple.” -Steve Jobs