Article by Tarek Faycal, Analyst at WAMDA

Even though Artificial Intelligence (AI) is developing at a blistering pace, outside academia and engineering, it remains a bit of a mystery. More often than not, a slightly intimidating one. Some people conflate AI with a dystopian machine-controlled future, such as that depicted in ‘The Matrix’, while others fear the proliferation of autonomous robots that wreak havoc like the Terminator or Robocop (although Dubai Police seem to have a handle on this last one).

In reality, AI is a field concerned with designing intelligent systems that exhibit traits associated with human intelligence. This includes understanding and speaking languages, solving various problems, and learning by example.
Globally, the industry attracted $26-$39 billion in investments last year. $20-$30 billion of this sum was in the form of internal corporate investment by global tech giants, 90 percent of which went to R&D, while the remaining 10 percent went to M&A. In 2016, AI startups received $6-$9 billion of investments in the form of venture capital, private equity, or other external funding. This was triple the amount reported in 2012.

The GITEX Future Stars event taking place this October will be further highlighting the AI industry’s latest advancements. The event that will be running from the 8th to the 12th in Dubai, is in collaboration with Wamda. Applications are still open.

Where is AI used?

Different sectors of AI are replicating specific aspects of the human intelligence in order to solve various problems. Below are examples of these sectors, along with MENA companies that have gained some renown for their work.

• Machine Learning (ML):

Concerned with giving computers the ability to learn without being programmed to do so, ML encompasses different ways of teaching machines to accomplish a task. Currently, the most commercially applicable approach has been ‘Supervised Learning’, which involves training the ML algorithm by using examples of input paired with matching output for tackling a specific problem. For example, a machine can be taught to detect cancer by showing it CT scans and their corresponding diagnoses. After many thousands of examples, it would learn to identify features that indicate a high probability of cancer. The current surge in AI applications is the direct result of developments in ML, including deep learning and reinforcement learning, which have enabled programs to perform increasingly complex tasks, such as defeating world champions in the game of Go and winning poker tournaments. This marks a significant milestone, as both short and long-term strategic thinking and even deception are required, often based on incomplete information. Furthermore, these ML techniques are now being put to use in other domains of AI, such as computer vision and autonomous vehicles. Neotic, is a platform that allows traders to test their strategies before deploying them in financial markets and provides relevant recommendations based on current and historical data.

• Computer vision:

Computer vision is a field that aims to help computers see and visually extract information from images or sequences of images. NAR, is a startup that’s currently developing AI software that would enable drones to autonomously patrol pipelines and send alerts when something needs attending to, as well as automatically generate reports.

• Autonomous vehicles:

Building on the success of recent ML advances, companies are now using AI to build the ‘brain’ of the self-driving car. Once equipped with such a brain, the car is taught proper driving techniques via interactions in a simulated environment coupled with multiple trial and error runs. However, there are several obstacles to overcome before autonomous vehicles can become the norm. The chief snag would be the lack of a total safety guarantee. Even if these vehicles make no mistakes 99.9 percent of the time, a 0.1 percent error rate would still pose a significant danger to passengers and others. Next Future Transportation Inc is developing modular autonomous pods that can join, detach, and even recharge while in motion. It partnered with the Dubai Road and Transport Authority and Careem to showcase the potential of this transportation method in GITEX 2016.

• Natural language processing (NLP):

This branch of AI is concerned with designing systems that process and understand human language. Arabot is an intelligent chatbot that uses NLP, among other techniques, to create easy-to-deploy chatbots for a variety of industries and instances.

• Speech recognition:

The translation of sound waves into readable languages is what speech recognition is all about. Significant strides have been made in recent years to make speech recognition software more practical. A studyconducted in 2016 at Stanford University showed that speech recognition in English was three times as fast as typing, and Google recently announced that its speech recognition technology achieved a word error rate of around five percent for US English. With such performances, it is not hard to see why this interface is becoming ubiquitous.

Votek Inc is a UAE-based speech recognition software company founded in 2014. It has collaborated with regional governments and companies to provide Arabic voice recognition services, in addition to ‘intelligent assistants’. Votek has also developed an educational toy, Loujee, that uses Arabic NLP and speech recognition to provide entertainment for children.

Consumer trust in AI is increasing

According to Accenture’s 2017 Digital Consumer Survey, 76 percent of users in the UAE are comfortable with AI customer service (the global rate stands at 62 percent). They even view it as more advantageous than interacting with humans. The advantages cited include availability at all times, a lower level of bias, and faster engagement.

In another survey, conducted by PwC, a majority of respondents in Saudi Arabia (66 percent), Qatar (65 percent), and the UAE (62 percent) revealed a willingness to replace doctors with AI and robots. In the case of healthcare, the primary motivators were speed and accuracy of diagnosis and treatment. However, a lack of trust in robots’ judgment when making consequential decisions as well as the absence of that irreplicable human touch were cited by 47 and 41 percent of respondents, respectively, as the main reasons for their reluctance to undergo some treatment at the hands of AI-backed machines.

The UAE takes the initiative

The UAE, specifically Dubai Future Accelerators (DFA), has already invested in a significant number of startups utilizing AI across different sectors, including QueNext, Avalon AI, Gramlabs, and Market IQ.

Another DFA-supported company, Comae Technologies, entered into a partnership with Dubai Police to develop AI-based forensics solutions.

Meanwhile, Cognit, the joint venture between IBM Watson and Mubadala Development, promises to bring the decision-making power of cognitive computing to businesses in the UAE.

AI developments are slowly seeping into various products and services across the globe. Seeing a wholesale adoption of this advancement will depend on the initiatives companies will take to adapt and integrate AI into their core services. It will also depend on customers’ receptiveness to the resulting significant changes.

View original article here: https://www.wamda.com/2017/08/ai-taking-roots-mena