Artificial Intelligence: How Close Are We?
2016 was a fantastic year for Artificial Intelligence (AI). A number of tech companies – including Amazon, Facebook, Google, Microsoft, Tesla Motors and Baidu – cashed in on their substantial investments in the field.
However, while there have been major advancements in AI in recent years – from autonomous vehicles to AI-enhanced household assistants – we’re still a good way off from developing machines that can match human cognitive function across the board.
What exactly is AI?
Many people associate the term “artificial intelligence” with the futuristic robots of science-fiction. But AI simply refers to intelligence originating from computers, and exists in many different devices, from lighting to self-driving cars.
There are three main categories of artificial intelligence – artificial narrow intelligence (ANI); artificial general intelligence (AGI); and artificial superintelligence (ASI).
Artificial narrow intelligence (ANI)
Artificial narrow intelligence, sometimes called weak AI, is an artificial cognitive system that can match human ability, but is limited to a specific function.
For example, autonomous vehicles, Google voice search, and robots at manufacturing plants all use ANI systems.
Although ANI is the most basic level of artificial intelligence, it’s paving the way for the development of computers with more advanced capabilities.
Artificial general intelligence (AGI)
Artificial general intelligence is what most people think of when they hear the term AI – a computer that can match human intelligence across the board.
We have made strides in machine learning – where a computer is trained to function using extensive sample data. However, it’s estimated that we’ll need anything from another five to 20 years to develop true AGI, where computers can perform any intellectual task that a human can.
Artificial superintelligence (ASI)
The most advanced form of AI is artificial superintelligence, which refers to machines with an intellectual capacity greater than that of humans. We’re likely to wait a number of decades for this level of AI to be achieved.
Where are we now?
In the past few years, ANI systems have been incorporated into everyday devices – from cellphones and computers to smart watches. For example, the map application on your cellphone uses ANI to navigate, and your email client uses ANI to filter spam.
Google search is one large ANI system, which uses sophisticated methods to determine the best results for your searches. There’s also a network of ANI systems online that share information with one another regarding your purchases, internet searches, and music preferences in order to generate recommendations for you.
ANI is now also commonplace in the military, in manufacturing and in the financial sector.
One major advancement in ANI that has been closely watched is the creation of autonomous vehicles, featuring ANI systems that can perceive and react to their environment. Google and Tesla are among the leaders of the pack. They believe we can expect self-driving vehicles to be available within the next few years.
The way forward
Computers can already easily perform a number of tasks that seem difficult to humans, such as solving mathematical equations, translating languages or playing chess. However, the difficulty in developing AGI is that, so far, computers can’t match people’s abilities of perception.
How do we then bring human-level intelligence to computers? One way of doing this would be to emulate the structure of the human brain.
For now, scientists have been able to successfully emulate the 302 total neurons in the brain of a flatworm, which is a far cry from the 100 billion neurons contained in a human brain. This makes human brain emulation seem like an impossibly difficult feat.
The solution to creating AGI may be to simulate evolution rather than mimic biology. In order to achieve this evolution, scientists will need to test computers at specific tasks, and then merge the programming of the top-achieving computers together to create a new (and hopefully smarter) computer.
What we can expect in the near future
For now, the focus is likely to remain on attempts at applying reinforced learning – where a computer “learns” through trial and error – in the production of self-driving vehicles and industrial robots.
Also, China is expected to become a major player as it moves away from copying Western models and strives for innovation in the fields of both AI and machine learning.
PM&A Consulting: meeting your networking needs
At PM&A Consulting, we specialise in providing small to medium Cape Town businesses with expert networking and IT security services, as well as leading hardware and software products.
Full artificial intelligence of the kind imagined in sci-fi novels might not be here yet, but we can provide practical, affordable ways to make computing easier, more powerful and more secure for your business.