The emergence of increasingly aware ‘cognitive computing’ implies a new era of technology and human collaboration. Google Now and Apple Siri lead an initial wave of intelligent ‘virtual assistants’ that can help you exit traffic early or choose a restaurant based on prior medical history; this ‘No channel experience’ has the potential to increase both brand agility and bottom-line efficiency. We are told that networked, AI-driven and cognitive technology could mean ‘white collar’ blues for our future skilled workforce; by 2020 we may employ machine before man. Embracing AI now will increase machine-based ‘human-like’ innovation at points of key interaction before your competition. Faster, more accurate and better decisions via ‘self-learning’ robotic software means leaner operations and better shareholder return but, don’t worry, a machine can only be so human can’t it?
Is AI really elementary my dear Watson?
Kurzweil famous predicted that by 2029, computers would pass the ‘Turing Test’ – the moment at which intelligent machine behaviour would be indistinguishable from that of a human; this prediction came before the arrival of the fax machine. He now predicts that by 2045, computers will be a billion times more powerful than all of the human brains on earth. That’s deep learning isn’t it?
Computers such as IBM Watson are on the threshold of commercially viable processing and indeed self-learning of semantic language content. Whilst their progression has not been stellar thus far, their ability to digest unstructured information at millions of times the speed of human cognition coupled with the emerging field of cognitive and quantum computing is narrowing the ‘context gap’ at a faster rate than ever before.
We stand at a tipping point comparable with the last industrial age. The next evolution in computing has the potential to create machines that will know the answer to our questions before we ask them (just like my wife!). This capability will be based on an insatiable ability to process enormous amounts of human and machine-generated log data in its raw and unstructured form to analytically derive context, meaning and perhaps most importantly of all, dynamic underlying relationships.
A large proportion of our effort in the current digital drive to business agility and indeed consumer and market responsiveness lies in the ‘drowning’ field of integration. This torrent of interface logic attempts to structure relationships that change too frequently; they are simply unstructurable.
What if in future, we left the analysis and interpretation of key data events inside and outside our organizations in the ‘virtual hands’ of cognitive computers powered by the crowd in the cloud?
Instead of today’s CIO focusing on ‘obsessive integration syndrome’ they would drive ‘federated intelligence’. IT would admit defeat in the fight for corporate control and structure of data and, subcontract this process to regulated machines that enhance core data against internal learning algorithms to adapt to the changing needs of the customer, employee, shareholder and regulator.
Are ‘white-collar’ futures in Jeopardy?
In 2011, IBM’s cognitive platform ‘Watson’, performed as a contestant on the US game show ‘Jeopardy’ and won. What is astounding is that its exploits were not coded by human engineers, but self-taught by reading Wikipedia – all of it.
This cognitive reasoning underpinned by semantic computing (or natural language processing) has the potential to automate patient diagnosis, research and development activities, product ideation cycles, Financial and Risk decisioning and supply-chain optimisation to name but a few potential candidates. It is even powering ‘personality profiling’ powered by big data and psychology.
The future for doctors, pharmacists, actuaries, quants, product designers, data scientists, auditors, recruiters and pilots will be very different than today. They will be pushed ‘higher up the information supply-chain’ to interpret, govern and action the change rather than participating in the journey.
A GE turbine alone generates more event data in one day than the entire global twitter feed.
This ‘data dark matter’ cannot be easily interpreted using current ‘structured’ and highly manual data processing techniques plus, the impact of the millennial generation and expanding ‘sharing economy’ is pushing future data ownership from the CEO back to the consumer in the crowd.
We see a future where machines facilitate strategic, economic and political decisions and accelerate a global process of creation, ideation and rationalisation. They will commoditise the acquisition, marshalling and interpretation of global thoughts, actions, events and sentiment from socially-enabled people, machines and sensors to enrich a globally accessible repository of ‘Open Data’.
The ‘most human’ of decisions in future may not require a pulse at all.
Early cognitive, deep learning and natural language capabilities have arrived. Your own ‘personal data sherlock’ may await you in these emerging technologies.
- Open Source – Apache Mahout, Caffe, Deep Learning 4J and Goggle’s Word2Vec
- Commercial products and services – Declara, Ersatz, Intelligent Artifacts, Numenta, Saffron and Watson
- Emerging API AI stores – AlchemyAPI, Cognitive Scale, IBM, Pivotal, 0xdata and Vicarious
Future IT service models will drive cloud-based adoption of increasingly cognitive ‘pick & mix’ application stores and APIs that will be combined to filter, analyse and deliver intelligent semantic context against global data events in a form that makes sense to your brand and employees.
We can now rent brain power by the hour and be in with the in-crowd.
After all, why bother to continually integrate your disparate systems against this torrent of global event data when, a cognitive ‘Enterprise Relationship Bus’ could summarise your daily status against potential risks and opportunities in direct linkage to your customers, products and services?
There are tangible real-world examples in place today from cognitive cooking through to personal diet advice, from improved heart disease diagnosis to reducing global crisis response times and from biodiversity and conservation through to faster recruitment screening; the possibilities are endless.
Business success lies in releasing control of corporate data assets, developing ‘Open Data’-based differential services and moving human knowledge workers up the client value chain.
Contribution by Simon Gratton
Part of Capgemini’s TechnoVision 2015 update series. See the overview here.