I have been fascinated with AI for as long as I can remember, an interest certainly influenced by my love of science fiction, robotics and that very human desire to replicate and improve through technology. Any student of 80’s sci-fi knows this ends with terminators and lots of back and forth through time correcting the decisions of mad mega-corps, this world is in my DNA.
We as a race have in a sense hardly moved on at all. Still we remain transfixed by the exhilarating dream of the possibilities of a brighter progressive world. No different really from that of the 18th century. But we are also still preoccupied by the consequences of that progress. Is it ethical, unbiased, simply right for us?
What is clear is that it is unstoppable, the genie is out of the bottle and the horse is already trampling over your cosy allotment. But should we be worried? If we take the time to understand it we might just see how our arguably non-artificial intelligence fits in.
We live in a world that is getting bigger and bigger and that includes data. To make sense of it all we deploy clever people with science, specifically mathematical algorithms. In that data can often be found patterns and connections related to human behaviour. Those there seemingly impenetrable hills of data can contain a goldmine for companies with the time and patience to dig. Big data engineers are the prospectors of our digital age and if they can find that rich vein for you then you have a potential competitive advantage in knowledge.
Advertising has become a rich area for new AI companies to flex their algorithms and do the maths. Programmatic advertising itself deploys machine learning to profile users, websites, viewability and calculate the relevance of a particular advert to that audience. The results are not always perfect but the machine is learning. Further to that, companies are appearing that claim to be able to tap into log files and further optimise delivery in the early stages of the supply chain. These algorithms allegedly reduce the cost per lead on certain platforms. Software-as-service and our increasing use of cloud-based tech has become a driver for innovation. We are ramping-up our development of AI systems to link and further service these evolving human behaviours and technologies.
Some techno preachers say that we are headed towards a singularity, a point in time and technology where anything becomes possible. Mechanical and biological intelligence can no longer be defined as separate entities and cloud-based intelligence permeates and informs everything we do. Any problem can be instantly queried, instantly solved. Sounds good right? You effectively have a cappuccino with chocolate sprinkles every day rather than on special occasions. Literally sweet!
Before we gamely plug ourselves into the future and fall into a sugary coma of awesomeness there are some things to consider. A certain tension between this relentless progress and the fear of what this may cost us. We lose control and have bad unhelpful AI by accident.
I am not going to go into too much detail on this other than to state that the conversation on the ethics and bias of AI is ongoing. Indeed the world renowned AI Professor Noel Sharkey along with Mary Wareham and Jody Williams along with various NGOs are in exhaustive talks with world governments to try to get legislation in place to stop us being taken out of the decision making process on AI weapon systems. It’s that serious.
When it comes to AI in advertising we are basically talking about multi linear regression algorithms and the like. These can be useful to find the trends which lead to better returns in the supply chain from buyers to publishers. It is a form of machine learning which may improve your ROI given some specific variables such as time, site, device and how changes effect revenue for example over time. AI can effectively help us manage complex situations that have a lot of variables, they provide the building blocks and focus but AI is no replacement however when it comes to creativity, deep insight and strategy.
Good AI should fit into our increasingly sophisticated and complex innovations to make them work for us, not get in the way.
So AI/machine learning/Deep Learning/Big Data can really help us understand human behaviour so we can make better decisions to reach them and with more and more relevant messages. AI can also help us organise and understand complex data so we can focus on the engineering and strategy. We can use that data and have more time for that cappuccino with chocolate sprinkles. I have however ordered a cyberdyne systems t-shirt off Amazon just in case future me needs to play the ‘staff’ card.
Snack Media are experts in the field of advanced programmatic advertising. Contact us if you would like to improve your digital advertising performance.
Author: Matt Tait, Operations Manager at Snack Media
For enquiries please email: matt.tait@snack-media.com
