Human assisted artificial intelligence
The Turing test explores one aspect of AI, the machines’ ability to demonstrate human-like intelligence. But whether these software are true AI, or human mimics, is debatable. The question soon becomes, ‘what is intelligence?’. Here we enter into the realm of the esoteric, of philosophy, and of academic exploration; and AI diffuses into a limitless futuristic intention. So a more practical question, a more commercially useful question is, what can we use our current AI achievements for?
The rising use of machinery during the Industrial Revolution replaced the most arduous and simplistic manual labour. Going forward, the rise of Artificial Intelligence systems is replacing the most mundane intellectual tasks, freeing up a new labour force to transfer their uniquely human intelligence to more interesting endeavours. This form of synergy between humans and AI is often called Human-Assisted Artificial Intelligence.
We visited DigitalGenius at Level39 for insight. DigitalGenius is utilising Human-Assisted Artificial Intelligence to automate customer service conversations for global financial institutions.
Imagine you’re on a website, trying to find specific information. If you can’t immediately find it, you’ll either give up or start using a system to narrow down your query. FAQs will give a range of possibilities, based on keywords. A search engine will do the same. Although the information is available - somewhere - it can be difficult for a user to access the right answers to their specific queries.
The next advancement in brand-to-consumer communication is Human-Assisted Artificial Intelligence. It’s a new way to engage with brands; instead of a call centre, a Human-Assisted Artificial Intelligence helps brands better communicate with their customers by combining Artificial Intelligence to answer simple, repetitive questions, leaving human intelligence to concentrate on more sophisticated queries.
Without AI, an agent may have as many as a dozen simultaneous conversations, with varying degrees of success. Even at a ratio of 12:1, it’s hardly scalable for a million users. And it’s not very popular; it has time lag and returns obviously pre-prepared responses. It feels like a patched up solution.
Agents are more efficient, however, when augmented with a conversational system that understands and responds to users’ queries in natural language.
When taken into a commercial context, this becomes a clear brief. An intelligent Human-Assisted AI not only automates repetitive questions, it understands the brand and can hold an intelligent conversation with its customers.
Traditional systems can be clunky and unhelpful; a Human-Assisted AI knows what it knows and knows what it does not know. When the AI recognises the customer’s query is out of scope, it seamlessly passes to a human who can briefly intervene, then pass it back again.
Is this true AI? Or perhaps a hybrid Humanistic Interface? It’s definitely a more intelligent design. When it actively learns from its experiences, and is able to write extensions of its own code as a result, then perhaps we can call it a truly intelligent machine, comparable to its organic counterparts. Evolving within the context of simple commercial briefs can be seen as the first baby steps. Each one a notable achievement and definitive success.
And if it makes waiting in the call queue obsolete, it’s already earned its keep.
From Digital Genius