The Blog

Dear Robots, We Need Guidance, Not Predictions...

As much as we need robots, robots need us to get to the promised land of harmonious, fluid and genuinely helpful AI support. And right now we are the ones guilty of not investing in the relationship.

The dust has settled on a terrifying presidential election and for the second time this year a dizzying democratic hangover is setting in across the globe, despite contrary poll predictions.

Who could have predicted this!? We know the answer to this question is a 'what' as well as 'who', as experts and citizens turn to Alexa, Siri or more sophisticated AI platforms for our presidential election coverage and predictions. I wonder if they can help clean up the mess we've made?

It's not just the political sphere hearing the cries of 'I told you so' from our prophetic and learned machine friends; in October 2016 for the first time, artificial intelligence was used to predict the outcomes of cases heard at a major European court.

Despite this triumph of machine learning prowess, I can't help but think we've got the whole AI thing just plain wrong. We treat our AI infused devices like silicon slaves, summoned to their master at the drop of a hat to answer our whims and daily ponderings; from the trivial to the profound.

Hands up if you're guilty of any of the following:

  • What's the weather like tomorrow?
  • Should I have pizza for dinner?
  • Should I go on a date with Monica / Paul / Carla / James?
  • What is the meaning of life?
  • Who will win the election?

All fun and games but these use cases have a shelf life; as the novelty value fades and the dust collects, we're left feeling short-changed on our promised intelligent and automated ifuture. Why? Because we're missing the point of AI completely. We're not there yet. Which is just one reason why (irrespective of positive predictions and modelling) we won't be electing AI court judges any time soon.

We can't yet expect our devices to behave like informed and intuitive consultants who read our emotions and body language to understand exact sentiment and second guess us. Siri isn't your best friend, a trained political correspondent or your life coach.

As much as we need robots, robots need us to get to the promised land of harmonious, fluid and genuinely helpful AI support. And right now we are the ones guilty of not investing in the relationship.

The outcries of what happened at this week's presidential election can be heard ricocheting across the echo chamber of Clinton supporters and Trump doubters worldwide. The sound will dance malevolently in the eardrums of Brexit 'Remain' voters, who are all too familiar with the gulps of 'what have we done' rising in the throats of millions of Americans.

How we failed

This presidential election, perhaps more than any other, required a guided internet journey that AI was perfectly placed to deliver. With so much political content out there, AI is poised to help us sift through data, identify and serve the content we need most to make more informed decisions.

Beyond political pop quizzes, we should be taking a leaf from cutting edge consumer shopping experiences which, only now have begun to understand the need for guided internet journeys to support our decision making. The technologies currently available to segment, infer, recommend and automate powerful video content, based on our priorities need to be glued together to work effectively. This would eventually help us cast light on the facts that so often get lost in political rhetoric and Fox News reporting.

If women's rights are pivotal to you and there's a TV debate with important information on, this content should be served to you, whether you are tuned in on the day or not. Whilst watching this debate, a second screen experience delivering relevant and supplementary content would help viewers see beyond the political jousting and help them decide for themselves.

Apply theories from online consumer shopping and you start to build a picture of what a more intelligent decision making journey might look like for politics:

  1. Timely: People want to interact with content anywhere at any time and have micro moment experiences that build upon their existing knowledge base
  2. Discoverable: They crave guidance and context to assist in digesting different types of information, with differing perspectives
  3. Personal: They desire truly relevant content directed precisely to their needs
  4. Simplified: They will expect all interactions to be convenient and easy

If voters had better access to this type of smart technology we may have seen a different result, at least from the swing states.

The need for assistive AI is born from a human paradox.

  • There is an overwhelming amount of web content available
  • We crave more content than ever before to make informed decisions
  • We have less time than ever to choose and digest content

Having lost faith in people to do the right thing, here is an open letter and invitation to AI to set the record straight and start a new dialogue:

Dear AI,

I'm sorry we have treated you so badly to date. We didn't mean it, we just didn't know what we were doing. Things are a little out of hand. To apologise, I would like to invite you to dinner and drinks at my house to discuss how we can work together for a better future. I don't expect you to have all the answers but that's okay, I can help you. If you promise to help cut through the clutter and guide me to the content I'm looking for, at the right time to make better decisions, I promise to teach you about me and other humans. I'll be serving hot chips and Silicone soft scoop for desert. Hope you can make it and maybe next time, together, we can avoid this awful mess we're in.

yours sincerely,


Picture used with permission

Popular in the Community