How To Better Apply Artificial Intelligence In Our World
The world of Artificial Intelligence (AI) is a complete and utter mess. This is, actually, a very good sign. AI is also quite misunderstood outside of the sector itself. The words Artificial Intelligence are an umbrella term. They cover Machine Learning (ML), Natural Language Processing (NLP), Neural Networks and more. AI is all around us today, from our smartphones and speakers to crime fighting, the legal world, healthcare, marketing and manufacturing. But AI is also very narrow. It’s good and doing very specific, formulaic work, although that’s changing. Very slowly.
But AI remains messy with a patchwork of solutions. Some good, many failing. Ethics in AI, such as dealing with racial and gender bias, still lag adoption by industry. And government. Not many independent AI companies are really making a lot of money. There’s a reason we have yet to see a big M&A type of event with an AI company. Although many get scooped up by tech giants like Apple, Google or Amazon.
AI is Trying to do too many Big Things
I speak with AI companies a fair bit. I also speak with companies implementing AI solutions as well. Sometimes such integrations go fairly well, most of the time they don’t. Usually because of poor change management within the implementing company and a lack of understanding their workplace culture.
The culture of the majority of AI companies I’ve spoken too is one of trying to do big things too fast. As many of them are startups or early stage, maybe into a B series raise, they’re pressured by investors, shareholders, and the perceived need to do something big. This is unfortunate. Not all innovation can be forced.
When we see news headlines about impressive, huge AI advances, one can note that the AI solution being reported on is still often in the research phase. With caveats such as “one day the company hopes to…” or “In the very near future…” with far less instances of the AI product changing an entire industry.
The AI systems that are working very well, such as the assistant in our smartphones, have been around for several years, they are highly focused and subtle. And that may be where we need to focus AI for the near future. Subtle, almost invisible, applications.
Subtle Artificial Intelligence
Rather than taking moonshots with AI and trying to do to many big things at once, perhaps we could step back and focus on solving more small problems and tossing in a side dish of critical thinking to look at the longer term implications of what is being solved for.
One failure I’ve seen with AI startups is a lack of critical thinking, or the ability to see an AI solution within the bigger context in which it may operate. It is a reductionist approach, rationalisation brought on by the tech industry’s over-rationalisation and being addicted to problem solving and avoiding much in the way of critical thinking. It is in large part why we have racial and gender bias in AI. Critical thinking enables a more holistic approach.
Some AI companies are taking a more holistic approach, focusing on their area of expertise, but seeing the complex systems within which an AI solution must work in order to succeed. They’re quiet, working within their industry niche. They know their AI solution can apply to many vertical markets, but they focus on getting it right in one before moving across to another. This is a highly strategic play. It is also likely to produce solutions that work much better and aren’t vapour ware.
This subtle approach unfortunately, doesn’t quite fit with many investors and board directors today. Not in a world where rapid builds and quick exits are the flavour of the week. Such an approach may work well (but also isn’t right) in the SaaS world of digital products. It doesn’t work in the world of Artificial Intelligence.
AI can have profound rippling effects across a company, a culture and society. AI needs to be seen more holistically and how it fits into our sociocultural systems.
AI is not now, nor will it ever, take all of our jobs. Yes, it’s the darling of the art world. It won’t replace the desire for human art, to think it will is to not use critical thinking. Nor will AI replace literary genius. Yes, AI may write sports articles and lousy marketing copy, but it won’t replace the genius of the human mind.
What AI will do, as it evolves, is become a sidekick for humanity. An AI tool, Perceptron, was recently launched that solves complex math problems and translates over 200 languages. But we’re not quite sure what the true value to industry or society is yet. Imagine what we can do when we combine AI with quantum computing and toss the complex problem of climate change at them?
Let’s Stay Narrow for a While
For now and for perhaps a very long time, AI will stay what it is today; narrow. General AI is the term used to describe a fully conscious AI that can interact with humans in a human manner. This is impossible. It cannot be like humans because it is a machine. It is not human. We don’t even know ourselves what consciousness is. AI may become super intelligent, that does not mean it is conscious. Ascribing any sense of “humanity” to a machine is a failure of imagination.
But if we stay narrow, really focus on making AI work at doing subtle jobs, or being our sidekick, then that’s where real money can be made and we can truly find ways for AI to benefit human societies and cultures. AI will have deeper value and we can solve for much of the biases inherent in AI today.
