It’s too early to worry about a sentient AI apocalypse. The reality is that we know very little about how the human brain works — which means we know even less about how to build a computer that works just like the human brain. For very specific tasks, AI tends to make rapid progress until it matches human-level performance; then progress tends to slow down. So despite fears of an AI dystopia, the technology is still very limited compared to human intelligence.
A more practical problem in AI is figuring out good ways for engineers and product managers to communicate a shared vision for how to actually use AI in the enterprise.
What exactly is AI?
Here’s a working definition of AI:
AI is when a machine performs a task that human beings find interesting, useful and difficult to do. Your system is artificially intelligent if, for example, machine-learning algorithms infer a customer’s need and recommend a solution.
To put our definition in more context, the chart below describes the progression from analytics to AI, using banking as an example.
How to use AI
For AI to be useful, it must be customized to your specific business context. The scarcest resource in AI today is talent — people who can apply AI to reshape your enterprise for better outcomes. Your vision should be uniquely yours and specific to your organization, strategy and existing technologies.
Here’s what it looks like in various industries when you start to reimagine your company based on the innovation that AI can bring:
- Artificial intelligence in banking: How to be more trustworthy, efficient and helpful
- Artificial intelligence in healthcare: How to cut waste, enhance care and improve lives
- Artificial intelligence in insurance: How to make insurance more personal, affordable and adaptable
- Artificial intelligence in manufacturing: How to run longer, run better and keep relevant
- Artificial intelligence in government: How to be more efficient, effective and responsive
- Artificial intelligence in travel and transportation: How to take care of the fleet, the business and the passenger
Each link above includes a diagram, illustrated below, that shows the AI possibilities for that industry.
Start small and experiment – before your competitor does
Remember that the nature of applying AI is experimental. You won’t know ahead of time which technologies and applications will be the most useful. Avoid biting off large AI projects all at once. Instead, run small experiments that make it easy for you to recover from mistakes. Create a portfolio of hypotheses about what you think might make a real difference. Test those hypotheses using small experiments. Learn and adjust as you go.
Remember, also, that AI can mean competitive advantage. It doesn’t make sense to wait until your entire company has mastered more basic analytics before taking on AI. Some areas of your business may be ripe for AI today. Find those areas and make them as smart as possible as quickly as possible.
The real danger of AI
A rule of thumb proposed by AI pioneer Andrew Ng is that anything that the typical human being can do without a second of thought is a good candidate for automation using AI. This puts a lot of jobs squarely in the cross hairs of AI automation. Worse, the people currently doing those jobs are probably unaware.
AI will most likely cause the problem of job displacement. A big challenge is how to create educational opportunities for those pushed out of their jobs by AI. How do we provide the training that gives them the best shot at finding meaningful work? Companies will need to take an honest look at impending labor shifts and balance AI innovation with re-training programs for their people.