The Power Of Collective Intelligence

The Power Of Collective Intelligence

Philip Hammond’s budget included measures to accelerate UK business’s application of a range of new technologies, including Artificial Intelligence, which benefits from a new £75 million fund. But there are already steps any business can take today to optimise its approach for the eventual – and inevitable – application of AI to its workforce.

The world’s best chess player currently is neither human nor machine. Instead, it goes by the name of Intagrand and is a combination of multiple AIs and people - all working together.

The most important kinds of intelligence in the foreseeable future won’t be AI working alone but, rather, a collective intelligence that includes both people and computers. And this is where there is a major opportunity currently untapped by many organizations

If we look beyond the deadlines concerning AI - bypassing the hype, ambiguity and fear-mongering - the real question organisations should now be asking is: How do we connect people and computers so that - together - they act more intelligently than any person(s) or computer has done before?

When considering how AI can complement human intelligence in their business, a useful approach is to split the human/AI dynamic out into four specific relationships to understand each before considering the combined potential of the whole.

First, there is the use of AI as a tool, with the computer performing intelligent tasks while people monitoring every step. A good example is the intelligent credit scoring algorithms that help human loan officers decide who they should give loans to. This dynamic is the most common so far.

Second is where AI becomes an assistant to employees. Unlike a AI tool, an AI assistant can work without direct human attention. KLM’s customer services chatbot, for example, dynamically creates suggested replies to customer queries but a human customer services representative then decides to use or override the AI’s recommendations.

Third is a peer model with the AI broadly performing the same task as your employees, but variants of the task that the AI is incapable of solving get escalated to humans. Lemonade Insurance use this approach with the AI instantly settling insurance claims that are within a prescribed set of parameters but, if anomalies are identified, a claim gets raised to a human adjustor for review.

Lastly, there is the use of AI in management tasks. Uber and Lyft do this with drivers working to an AI manager that dictates how much they’ll be paid, mandates performance levels and sets schedules. At Cogito, the AI coaches customer service agents to improve their effectiveness when talking with customers by analysing factors including speech patterns, pauses and vocal strain.

As well as considering how we can divide tasks between humans and AI, it is also important to consider how these systems can improve over time with the benefit of experience.

Such improvement typically occurs in one of three ways. First when AI makes employees better - as Cogito does. Second, when employees make AI better – as Google does, with its programmers constantly tweaking the search algorithm to give a better experience. Finally, when an algorithm learns from its own experiences – as KLM’s chatbot does each time employees corrects AI-generated suggestions.

In humans, there are different types of human intelligence - academic knowledge, emotional intelligence, and so on. In AI there are different types of intelligence, too, and the way these all work together needs to be considered carefully.

As the Chancellor works to stimulate AI’s use in UK business, examples like Instagrand should act as important signposts to show that future success will come from the combinations of many differing kinds of human and AI - all creating a new kind of collective intelligence.

Tariq Khan is Director of Interactive at TMW Unlimited

Close

What's Hot