AI: The Silver Bullet For All Your Business Problems?
In opinion / By Mark Flynn / 22 October 2018
AI will revolutionise so many areas of our professional and personal lives. AI is the centrepiece of the fourth industrial revolution but quite frankly, right now, it is the most over hyped trend on the face of the planet and we need to get real about what it can really do.
Everyone is talking about AI or claiming their product is the best because it has "AI capabilities". The truth is however, that AI is never going to be a silver bullet to make all our problems disappear.
We all need to realise that AI is not some magic invention that will allow robots to compete with the human brain.
Yes, it can do some amazing things such as the 100% correct diagnosis of eye related illness, but these results take time and effort of a dedicated team and the results are very narrowly focused on this one medical condition.
AI is very task orientated, offering powerful solutions that drive autonomous cars, detect cancer tumours and allow PayPal to detect possible fraud.
However, the AI assistant in the autonomous car cannot detect the difference between a cat and an Elephant.
Put another way, Mo Farah will never beat Usain Bolt at the 100 metres, neither will Usain beat Mo at 10k because these world-class athletes are built for and trained all their lives for very specific, and very different sports.
So to successfully implement AI, you've got to go narrow, consider what it is exactly that you want your AI system to do. Every AI solution needs to be focused around a core task.
This specialisation will dictate the kind of data you’ll need to feed into the AI. Bear in mind that this will require a lot of heavy lifting before you'll see a return on any AI investment.
For example, AI will require a high degree of integration between your key business systems in order to build a single source of the truth that funnels up into your AI solution.
You've also got to ensure the data is organised in such a way that the AI can understand and cross-reference that data with other pieces of information as needed.
Now a lot of people assume that an AI won’t have a bias but the truth is the AI system is only as good as the data it is fed. Another spin on rubbish in, equals rubbish out. For example, Amazon had to ditch its 'sexist AI' recruiting tool that showed bias against women. In this case, the tool learned to favour men over women after being trained on past CVs written mostly by male candidates.
The last step before rolling out your AI enabled solution is to give it a test run. Real world testing always brings out flaws in any technology. A good example of this is Microsoft infamous chatbot, Tay.
Tay was a chatbot made by Microsoft used on Twitter until a prankster managed to feed Tay bad data, which turned the bot into a very vocal sexist who swore a lot on a public forum until it was hurriedly turned off.
So in summary, AI should not be seen as a revolution, instead a journey of hard yards that might actually start without any AI input!
A lot of companies will make a rash decision to use AI if they hear their competitor has already done so, however this is often a bad idea.
Adding AI isn’t going to instantly make your business better, and it is often best to release something without making use of AI.
Yes, AI is exciting but make sure you look beyond the froth and stick to your well-grounded business principles in order to get the best out of it.
For more information on anything in this article, please contact me.