People want voice assistants to be more personable – but what if they were actually human? In Amazon’s Super Bowl spot, celebs are brought in to help when Alexa loses her voice. The ensuing chaos shows how, while people love chatting with their AI, it can’t be at the expense of its efficiency.

Worldwide spending on robotics and related services is predicted to reach $135.4 billion by 2019 – but one thing hindering AI's progress is the 'uncanny valley'. This describes the unsettling feeling one experiences when a robot looks or acts too much like a human.

A global survey of Quartz readers has found that 79% of people are very or somewhat familiar with AI technology. But while they're than happy for it to lend a helping hand with their health, finances, or shopping, they’re not willing to give it control over their everyday decisions.

Over 70% of people fear AI in some way, with 24% worried it will take over the world. Canvas8 spoke to Tom Chatfield, author of Critical Thinking, to find out how we can create AI that won’t jeopardise our liberty, and to understand how human biases influence machine programming.