"It is the embodiment of human intelligence and there is really nothing more at present than a thin veneer of that – we simply don’t have the computing power to build a neural network like the human brain," Nathan Lowe, managing director, ASI Solutions, said in an interview.
“AI is something that performs a specific set of tasks in a human-like way. Human intelligence has creativity, innovation, sentience, morals, gut feel, is situation aware, bias, feelings, and has learned from its mistakes. AI like this is not going to happen for a while,” he said.
Lowe has been with ASI since 1999 and became managing director in May 2015.
How has the world been introduced to AI?
In the 60s and 70s it was Lost in Space’s robot, Star Trek’s computer and HAL from Kubrick’s 2001 A space odyssey. But these were still largely seen as computing devices. Although today ask Siri what she wants to be when she grows up? “I want to be the computer on Star Trek.”
In 2011, TV game-show Jeopardy had two former champions Ken Jennings and Brad Rutter playing against a third, an Artificial Intelligence type of computer system called Watson created by IBM. They went head-to-hard-drive and were solidly beaten by Watson showcasing IBM's expertise in advanced science and computing. That was the first real demonstration of the potential of AI.
Late last year, Uber users in San Francisco hailed a self-driving car to pick them up and take them to their destination. The cars were fitted with sensors and cameras, and through Machine Learning, were taught to steer, accelerate, brake and change lanes.
Where is AI/ML heading?
There is a difference between artificial uintelligence and machine learning.
True AI has the potential to create far more impact than the invention of the personal computer and the spread of mobile phones. It is the embodiment of human intelligence and there is really nothing more at present than a thin veneer of that – we simply don’t have the computing power to build a neural network like the human brain.
So let’s confine the advances to the reality of ML where this technology has moved us forward in technology stakes faster than we have ever imagined. Rather than trying to embody (program) machines with everything they need to know upfront, we want to allow them to learn from their observations of the world.
Everything in the world is now powered by a machine. Through the advancement of technology, engineers today can develop a machine that, within codified parameters, can learn on its own. This is Machine Learning. It seeks to create predictive models and algorithms. Giving computers the ability to carry out tasks without being explicitly programmed.
What are some examples of that?
Examples of machine learning we use on a day-to-day basis are Google search engines, recommendations from Amazon, Netflix, and YouTube, and even suggested friends on Facebook. Another way ML has enmeshed itself in our lives is through face recognition. At the airport, you are staring at the camera ahead, feet planted exactly as marked, waiting for the gates to open.
Engineers can implement machine learning with a basic level artificial intelligence, which consists of systems that enable computers to perform intelligent human tasks without being explicitly programmed. Examples of day-to-day common uses of artificial intelligence include Apple's Siri, computer games, fraud detection on credit cards, online customer support using Chatbots, and security surveillance.
Since computer hardware and software are rapidly advancing, Artificial Intelligence surpasses the capabilities of human experts, such as the win at Jeopardy or the Japanese robot that beat humans at market forecasting. There is great potential in the use of this technology, such as fewer errors in medical practice or fewer road accidents.
Artificial intelligence has features such as faster speed, integration of cameras, and precise speed recognition that allows machines to perform some tasks better than humans. But humans have not been replaced yet.
What are the complexities of ML?
The underlying technologies of machine learning are very complex. It has three parts – the model, the parameters, and the learner.
- The model is a system that makes the predictions and identifications.
- The parameters are the signals and factors the model uses to make its decisions
- The learner is the system that adjusts the parameters by looking at the differences, the predictions, and the actual outcomes.
What are the complexities of AI?
Artificial intelligence is vastly more complex. Each program depends greatly on the purpose of the product. They all have three components – data structure, inputs and outputs, and learning systems.
There are two types of data structures needed, one for long-term storage and the other for short term storage. Inputs and Outputs are the core sources of data. Some examples of Inputs are sensors and downloaded data. An associate of Learning System is the most crucial component of the Artificial Intelligence system. This presents us with a Learning System of the machine and tests boundaries. It gives the ability to perceive and learn new information and allows for cooperation and social intelligence upon human interaction.
Will AI be replacing us anytime soon?
Although machine learning and artificial intelligence are useful technologies, it is a long way off before they pose certain ethical dilemmas — the rise of the machines leading to subrogation of the humans — or more recently the implied threat that they will take our jobs.
By 2034, one analysis concluded that 47% of all jobs in the US could become automated, which means robots could take over human employment. This could be appealing to companies since robots do not require salaries, toilet breaks and do the job perfectly every time.
However, some in the tech world think Artificial Intelligence could allow people to enjoy their lives and use the intelligence as an enabler to accomplish more. To eliminate tedious and repetitive tasks and allow more time on creative and other fun endeavours, but still, have an income from other pursuits.
Many in the world of artificial intelligence and machine learning are wondering what happens if the machine fails? Who would be at fault, the programmers or the end-users? Since machines do not have advanced social intelligence (yet), how would they make complex and moral decisions?
As Max Tegmark, the president of the Future of Life Institute, said, "Everything we love about civilisation is a product of intelligence, so amplifying our human intelligence with artificial intelligence has the potential of helping civilisation flourish like never before – as long as we manage to keep the technology beneficial."
What about the Turing test – “let's make it conversational”?
Conversational systems are the voice interfaces many science fiction writers and technologists had only dreamed of. However, thanks to the advances in artificial intelligence and machine learning, voice-operated conversational systems have become more practical. There are many platforms, such as Apple's Siri, Microsoft's Cortana, Amazon's Alexa and Amazon's Echo. These voice interfaces can translate voice into search commands.
They also have the skills to manage song playlists, shopping lists and to look up information quickly. The future of these conversational systems is to control appliances, which is happening already with some connected white goods.
In business, these conversational voice interfaces — chatbots — can simplify business practices, where it can enable users and systems to have meaningful interactions. Companies are always finding new and innovating ways to increase the brand-to-consumer communication. There are new touch points with consumers that are relevant, highly personal, and conversational.
Powered by a combination of machine learning, natural language processing, and live operators, retailers and some tech firms are extending the conversational systems with chatbots. They are to provide customer service, sales support, and other commerce-related functions.
With the popularity of mobile messaging, voice-operated conversational systems and the advances in artificial intelligence and machine learning, the new generation of tools can enable companies/brands to communicate with customers faster, better and cheaper.