Have you read ‘The UX of AI’ by Josh Lovejoy?
The Staff Interaction Designer at Google talks about a lot of things in this rather interesting, certainly intimidating article of his. But, what stands out from all his dialogues, what I like best, is this one.
‘As long as your attempts are not aligned with a particular human need, you’d end up building something powerful only to realise that it solves a problem that either doesn’t matter or doesn’t exist.’
That’s the gist of AI design, right? Every stage in the process demands innovation. And yet, it can’t figure out its motive without human help. Base it on theory, and it will never solve a genuine, practical problem.
That’s why human-centred AI should be one of your prime concerns.
The Enthusiasm for AI Is Not Enough; It Must Also Be Morphed According to Human Concerns
In the early 2000s, a computer would have barely recognised sharp edges in a photograph.
Today, we have big data, neural network algorithms, powerful hardware, exceptionally improving computing strength. AI is no longer a niche. It’s a differentiator that is on its way to completely revolutionise several industries in the upcoming years.
And yet, if we refuse to let AI develop with human concerns, we’d end up with highly intelligent machines that have close to no practical usage.
The Need for a Human-Centered AI- Why?
1. The Future Artificial Intelligence Designs Must Understand Human Perception to Be of Appropriate Assistance
Aligning AI with the human thought process and its complete scope is a herculean task, no doubt there! It’d require reconnecting the developing tech with other fields, including but not restricted to psychology, cognitive science, computer vision, and sociology.
But, unless the AI software learns to perceive things as a human would, they wouldn’t understand the problems from our perspective. How do you then expect them to create long-term solutions?
2. We Must Define Artificial Intelligence Such That It Enhances Humans
Remember, the goal isn’t 100% automation. Instead, we need a combination of specialised hardware and wise software. That way, we’d get to strengthen and expand our capacities while ensuring that no human error, like distraction caused from fatigue, affects the output of a certain action.
Think of the changes such an AI machine would bring to fields like surgery, senior caretaking, etc.
3. While We Can Not Entirely Eliminate the Job-Displacement Threat, We Must Guide the AI Development at Each Stage to Minimise the Damage
Computers do not realise what they are doing. They are mind tools, as Jim Guszcza, Chief Data Scientist, Deloitte Consulting US, puts it.
If a machine is biased against underrepresented communities or doesn’t consider the privacy rights of any individual important, it’s a reflection of their design flaws. Sure, they will work fine on auto-pilot most of the times. But, there will be scenarios like the above where a certain ambiguity or exception will need human intervention to do justice to the case.
That, in the long-run, will help create applications of artificial intelligence that are helpful to the humans as well as least dangerous to their prospects.
The Ultimate Goal is Balanced Profit for the Societies and Economies: That’s Why We Need Human-Centric AI Approaches Now More than Ever
If we consider human psychology and needs, incorporating as much as possible in AI designs, we will be more likely to create something that not only offers better economic yield but also is more socially acceptable.
Keywords: AI design, human-centered AI, neural network algorithms, artificial intelligence future, AI software, define artificial intelligence, AI machine