Reliable, tireless and clever: our new digital friends

Artificial intelligence is becoming more intelligent, personal and predictive, but are we ready to hand over decision-making to machines?

mobile ai voice

2016 may have been the year of “peak keyboard”, where we moved beyond just tapping instructions into a keyboard, and started to communicate with machines in a more intuitive and human way, using voice and gesture. Digital assistants like Amazon’s Echo and Google’s Home Speaker, powered by Artificial Intelligence (AI), will allow us to chat with the Internet.

The Internet will become a conversation, no longer just a reference tool that we use to visit pages to look up information. AI assistants will tell us what time the restaurant opens today, as opposed to taking us to a static snapshot of opening times on a webpage. AI will be able to “close the deal” – book your train tickets, add toothpaste to your Amazon basket and check out, turn on the lights in your home and tell you how much petrol there is in your car.

ai in numbers

AI within these devices will also become predictive – it will book your Uber when you leave the lift at your office. They will be able to learn not just from your movements and instructions, but from the behavioural patterns of every single user on the platform. This is how AI scales so quickly – it learns not just from its own mistakes, but also from the mistakes of every other agent interfacing with humans.

The focus will no longer be on the devices we own, but instead on the AI assistant that follows us everywhere, accessed from the car, the office, at home, through many smart interfaces. It’s the start of what some are calling pervasive computing. These devices are always listening, they have to be “always on” to work, so we have to ask if we are ok trading convenience for privacy.

ai devices
As data becomes ubiquitous, we’re likely to transfer significant authority to AI, as the processing of data becomes too much for our wet, organic brains to manage.

We’re moving from a mobile-first world to an AI-first world. AI will become like a commodity, just like electricity, or wifi, which we will point at whatever problems challenge us, from supply chain management, to customer behaviour, to the origins of the universe.

As data becomes ubiquitous, we’re likely to transfer significant authority to AI, as the processing of data becomes too much for our wet, organic brains to manage. Access to almost unlimited intelligence will cause us to radically rethink how we ask questions – what would you ask with unlimited intelligence? Framing those questions will become a valuable skill.

To me, the real potential for AI is not about just automating what we can already do today, but pointing intelligence at things we can’t already solve. We face some very complex global challenges, from pandemic management to climate change to migration. Some argue we should design AI to address these problems, as humans have so far been unable to organise ourselves in time.

AI is not neutral, and is embedded with the biases of those who are programming it. Massive ethical and philosophical questions will likely define the coming decades, as we grapple with the digitisation of decision making: decisions that will have life and death implications.