July 19, 2018
Share this post:
Author: Chris Howard | IBM ANZ Watson Data Platform & DIP Practice Leader
A before I, except before C
Many of us will remember the mnemonic rule of thumb ‘I before E, except after C’ from our spelling drills at school. In today’s world of Artificial Intelligence and Cognitive computing, I would like to propose a somewhat modified alternative – ‘A before I, except before C’. Let me translate – in this case ‘C’ is for Cognitive (sounds very Sesame Street), with AI obviously referring to Artificial Intelligence. But in the rule above, I am arguing that before ‘C’, we should be looking to IA and not AI as the natural order of things. IA in this case, refers to Information Architecture, a foundational requirement necessary before any organisation begins to truly exploit the power of Cognitive systems and AI.
The phrase ‘information architecture’ appears to have been coined, or at least brought to wide attention, by Richard Saul Wurman (the creator of the TED conference) in the 1970’s. Interestingly a quick Google Trends search will reveal that this term has actually declined in use over the last decade, in stark contrast to the use of the term AI. Is this because IA just isn’t trendy, or are we really ignoring the fundamental need for information architecture?
Over the last decade, there has been a significant shift in the data management needs of many organisations as they look to tackle the 3 original V’s of Big Data (volume, variety and velocity). As the number and type of data source increases, it is the 4th V (veracity or truthfulness of data) that starts to present the greatest challenge.
Source: IBM Big Data Hub
So how do we ensure the truthfulness of your data – or in other words, how to have confidence in your decision making? Reducing data veracity and building data confidence (being able to rely on your data for the purpose of business decision making) is about understanding the lineage of the data – starting with its original source and understanding the journey it has taken (where has it landed, what transformations has undergone, and by whom). There are numerous definitions of Information Architecture, but one of the simplest comes from a 2011 whitepaper from Sybase, describing IA as the “holistic view on the flow of information in an enterprise, including the effects of the processes that act upon the data.” – sound familiar?
“Garbage in, garbage out!”
In computing terms, this phrase refers to the idea that incorrect or poor-quality input will produce faulty output. This too can be said for artificial intelligence. It is not just poor-quality data we should be worried about, but also low confidence data (untrustworthy data). We could argue that untrustworthy data as an input will produce untrustworthy output. Surely we should be striving for not just high quality, but high trust in the data that we use to prime the AI engines of the future. Unfortunately, many organisations are still taking too siloed an approach when analysing their data and ignoring the need to govern this valuable and strategic corporate asset. So before looking to AI and the enormous potential, it can deliver, take a moment to consider whether you are falling into the trap of “Garbage in, garbage out!”
Trusted data combined with the power of AI can radically shift your organisation’s ability to make better business decisions, differentiate against your competition and allows you to continue to remain relevant in the marketplace. BUT you need to be well equipped before you start your journey into AI – this begins with robust information architecture. Interested in knowing more? See this Ovum whitepaper on IBM’s Digital Insights Platform. (PDF, 122KB)
Author: Chris Howard
| IBM ANZ Watson Data Platform & DIP Practice Leader