Ghana's regulator at crossroads over 4G spectrum

At a regulatory
crossroads

Ghana contemplates
4G spectrum.

Wednesday, Nov 21st

Data-driven AI will force society to adapt

Data-driven AI will force society to adapt

A future with Artificial Intelligence (AI) will be defined by two distinct approaches:  Narrow Artificial Intelligence (NAI) and Artificial General Intelligence (AGI), both of which will require a combination of disciplines like quantum mechanics, data science and machine learning to leverage the opportunities.

This is according to senior executives at global analytics focused multinational SAS who hosted the company's 2018 Analytics Experience in Milan this week.

Oliver Schabenberger, Executive Vice President and Chief Technology Officer at SAS explained that NAI is a form of AI that solves very specific tasks using purpose-built systems.

"By October 2018 this data-driven approach to AI allows us to create very powerful transformation. Transform the data about a credit card transaction into a statement about its legitimacy. These systems do not think, they are plain algorithms."

Schabenberger said the creation of general thinking machines with the capacity and aptitude of human intelligence is another approach. "We call that AGI. Its aspiration is to realise in hardware and software human level intelligence1. AGI can solve general problems, any task, it is not built to do anything."

Today's AI technology in and of itself cannot lead directly to a future of fully applied, widespread AI. "We need the imagination of a radically different approach and that imagination somewhere at the fusion of material science, biology, quantum mechanics and computer science... what will lead to artificial intelligence is the fact we are trying and internal success will eventually get us there."

Schabenberger added that data science is at the heart of AI's revolution, along with an improvement in algorithms.

"No, AI is not just hype. We are today building better algorithms. AI is not based on writing down rules, it is based on training models by processing data and during the training the system develops its own logic, it is programmed implicitly. The AI boom that we are experiencing is an analytics boom enabled by massive digital audience through digital transformation, increased connectivity and advanced computing. "

Both AGI and NAI are highly disruptive, said the SAS executive who underlined the impact of this tech influence on jobs.

"I believe one hundred percent of jobs will be affected by some form or another of automation And that's not necessarily a bad thing. Narrow AI is great at automating frequent high volume tasks that do not require creativity, contextual understanding or logic reasoning, and that's great. The good stuff is left for us, the stuff that depends uniquely human qualities. AI diagnoses a medical condition, we provide the care.... AI translates language, we interpret the written word."

Schabenberger envisages a future in which more jobs will change with AI than will go away entirely.

"The concept of a job market as we know it today does not hold water if we have AGI. The goal is not to save jobs, the goal is to save humans."

And with this comes the need to be proactive rather than reactive in better understanding the implications of technology.

"Advances in AI are running ahead of our ability to absorb and integrate the technology. We have autonomous cars but we lack the traffic control system and no clarity on how to handle decisions in the face of harm. We know how to use math to diagnose cancer, but we do not know who is responsible when the maths is wrong. Technology is so fundamentally transformative and has such deep implications when mistakes happen, being better at inventing stuff than using it, is a lousy strategy."

ALSO ON ITWEB AFRICA

Taxify gains traction in Africa Published on 23 February 2017

Company says its drivers on the continent make more trips than European counterparts and expansion is on the cards.

Algeria, Nigeria eager to complete trans-Saharan fibre project Published on 02 November 2018

Algeria section completed, with feasibility studies for Niger and Chad also finalised.

Arch Mobile, a valuable management tool in receiving stock Published on 20 November 2018

Spinnaker has developed the Arch Mobile (AM) application, an extension of Arch Retail. The application runs on smart devises over a WIFI network offering users the ability to perform designated functions and transactions in real time.