It’s a question that has preoccupied computer scientists, science fiction writers, and conspiracy theorists for years: is it possible that artificial intelligence will one-day rival human intelligence?
It is a thought-provoking question but for retailers, insurance firms and other businesses serious about using A.I. to gain a strategic advantage, it’s also the wrong question.
True A.I, based on reinforcement learning, not only rivals but far surpasses human intelligence in many different areas.
That doesn’t mean A.I. acts or feels like a human being, or in attempts to compete with living, breathing people. It means A.I. is a tool that can be used to do things that people simply can’t do.
That’s the strategic advantage, but amid the hype, it’s easy to lose sight of that.
Some so-called experts might tell you that the best way to get started with A.I. is to look for projects that leverage the technology in ways that pose little risk.
But it doesn’t make sense to deploy A.I. to tasks and activities that humans can easily do.
To be blunt, many A.I. products are “low-hanging” fruit that fails to leverage the power of A.I. to address complex problems and processes.
Here are a few examples:
- Driving: While self-driving vehicles are intriguing and could take bad drivers off the road, driving is something that human can easily do and many people like driving. A.I. has been applied to autonomous vehicles because there are a limited number of potential scenarios such as routes, the movement of other vehicles, and weather conditions; something that can be easily addressed with computing power and software.
- Customer service: Visit a Website and you’ll likely see a chatbot that promises to answer all your questions. Again, does A.I. really need to be applied to customer service? People are good at delivering customer service. They can communicate and be empathetic.
- Email: Gmail caused a stir recently when it started to automatically provide people with content for email responses. Rather than typing a sentence, Gmail delivers, saving you seconds of “work”. Yes, it’s a creative use of A.I. but auto-answers are likely not something that people requested or, in fact, really need
As important, there are many things that are bad for humanity that A.I. should not be applied to such as:
- A.I.-enabled military weapons. We’re talking about fully autonomous weapons using A.I. that are programmed to kill. These types of weapons could be devastating, particularly if they fall into the wrong hands.
- A.I. that causes widespread job losses. A.I. will impact the work currently done by people but jobs shouldn’t disappear. It’s interesting that the World Economic Forum believes that 58 million new jobs will be created due to advances in A.I.
- A.I. for personal relations. There is a growing interest in robots that can communicate with humans. But what are the ethical considerations around human/robot interactions? Can a “relationship” with a robot have the same emotional connection as a human? And how far should a human/robot relationship go?
- Software that sparks a human dopamine response that’s addictive such as social media and video games – Fortnite, anyone? Many software companies are working on “persuasive technology” that influences behavior and alters habits. This would be a huge waste of human productivity.
Now, that we’ve looked at things that A.I. shouldn’t be used for, let’s look at the sweet spots for A.I.
- A.I. should power technology that makes tens of thousands to millions of decisions per week. This massive amount of work is far beyond the capability of humans.
- Highly complex mathematical processes in which there are millions, if not billions, of potential variations. Think of the ripple effect in retailing. When a specific product is promoted, there is an impact on other products due to cannibalization and pantry loading. This is one of the ways that Daisy’s A.I.-powered technology makes it easier for grocery retailers to make better and faster decisions about the products to promote and how to price them, leading to higher sales and profits.
- Relentless and repetitive work that happens every day, every week without ever stopping. Jobs like these should be handled by A.I. because they are relatively simple and time-consuming. People should focus on doing work that involves critical thinking, imagination, strategy, and creativity. (Note: For more thoughts on future job skills that A.I. can’t do better than humans, check out Bernard Marr’s article in Forbes.
- Time-consuming and manual tasks that could easily be automated. Think about sales, for example. How much more could a salesperson achieve if tasks such as loading information into a CRM, scheduling meetings, completing RFPs and creating reports were done by A.I.-powered technology. It would give salespeople more time to develop relationships and deploy new and different strategies to nurture and close deals.
The power of reinforcement learning
We believe that one of the keys to successfully leverage A.I. is reinforcement learning, which not only looks at historical data but simulates models of what might happen in the future by incorporating the most recent information.
Reinforcement learning works with data on an order of magnitude that would leave the smartest people on Earth in its dust. We’re talking about billions of alternative scenarios assessed before delivering recommendation about what a business should do.
This is work our brains were never designed to do. It’s too complex, labor-intensive, and time-consuming. This is why companies should deploy A.I.
If businesses want to truly leverage A.I., the technology needs to be applied to tasks beyond human capability to drive efficiencies, productivity and, as important, ROI.
Let’s look beyond the low-hanging fruit in A.I. This is not a competition between humankind vs. machine. It’s a competition to see who will realize the value first, and emerge as the leader in their market.
To learn more about how Daisy’s A.I.-powered technology powers business decisions that drive higher sales and profits, contact us.