Can AI professionals overcome these obstacles?

The economy might be slow and global trade tensions high, but a strong technology always finds its way of survival. It has never been more accurate as for AI.


Artificial Intelligence is worming its way into people’s daily lives, and has already made a thrilling entry – with video analytics and facial recognition in surveillance systems, cybernetic limbs for disabled, processing multiple languages with natural language processing neural networks used in chatbots & translation software, AI-driven personal assistants like Alexa and Echo, and many others. These are just the beginning of all that can become of AI.


But the question is, can AI professionals survive AI’s biggest obstacles? No, it’s not about killer-robots, but those challenges are more practical and immediate in nature. Before exploring them, let’s look at the future AI has in hold.


The promise of AI


Accenture predicts,


By 2035, AI-powered push will provide as much as $14 trillion boost to the global economy.


Such is the potential of AI industry. This technological revolution is led by Industry 4.0 and aims to completely overhaul the way we live. Industry 4.0 is the catch-all term for implementation of big data, improved robotics, and AI systems. Past trends also show the future of AI is promising. Marc Carrel-Billiard, Global Senior Managing Director, Accenture Labs tells,


Just 5 years ago, AI bots could resolve one of every ten customer phone calls with success. Today, it’s 60%.


Let’s not get ahead of ourselves – The critics are calling… 


If the future is promising, where lies the problem.


As the world heads into 2020 and has had a test drive with AI, the problem areas are more prosaic in nature. That’s not to say there are no legitimate concerns about many unintended consequences of AI, but those of the likes of killer-robots are not so immediate and are yet to see the light of the day (for the AI tech to advance).  


Here are three criticisms leveled by AI practitioners and scholars, most of which according to these experts are permanent limitations of AI, very difficult to overcome for AI professionals.


  1. AI systems are not so intelligent.


A lot of work is needed to make it truly intelligent, says Carrel-Billiard. At present, the AI systems are not very intelligent. It might do a decent job telling you how to go about the fastest route, book your tickets, and solve your generic queries, but it still doesn’t have a wide array of knowledge and expertise.


AI technologies must run on many iterations – machine learning, image and voice recognition, natural language processing – to be highly specialized yet generic to solve problems independently. He notes, AI is trained to interpret certain data sets, but it cannot by itself contextualize the complex world we live in. It would need to have all those thousand iterations in place to be able to infer meaning of any situation. That is AI’s blind spot.


At present, AI is being a specialist used for specific tasks, but for it to scale it should become a generalist.


  1. Deep learning is no replacement for deep understanding.


Another attack that AI professionals should take a deep look comes from Gray Marcus, the professor of Psychology and Neural Science at New York University. He says, deep learning (DL) is the subset of AI. Well, that’s no news, but what is, he says, DL hailed as the tech with an ability to make sense of huge datasets with little to no oversight of humans, is a misnomer or inaccurate. It is good at performing narrowly focused tasks, but he seriously doubts the potential of AI to revolutionize, say transportation through self-driving cars, and medicine through self-analyzing MRI scans for cancerous growths.


He asks, “(What are) the number of radiologists who have been replaced by deep-learning systems.” “Zero.”


  1. They are only as good as rabbits and lack depth.


AI critics say that for AI systems to be truly transformational and effective, they should be designed to be accountable, transparent and free of bias. At present, AI systems are not only learning how to perform tasks, but when it comes to decision making, they are also imbibing human biases; the same way a child mimics bad behavior of parents. If not corrected, critics believe AI systems will just be fast rabbits for humans limited to be used for standalone tasks.


Skilling up for new age


For AI professionals to combat these obstacles mentioned by experts, skilling and reskilling is urgent. An interim report on the Work of Future by MIT taskforce reiterates,


Technology may destroy jobs in the short run but will create better and new ones in the longer term.


While the job losses would not be instant, the labor market must prepare themselves. Young workforce would be at the instant risk of these changes. Of required AI skills, data analysts are being trained in adequate numbers, but engineers who can fully understand and do justice to AI systems are not enough.


The need for skilled AI engineers and AI professionals is in fact more critical in startups. Since legacy systems are always slow to change, startups are exploring the untrodden terrains to build AI and ML centered technologies.


Industry certification bodies, such as Artificial Intelligence Board of America (ARTIBA) are at the forefront of ongoing reskilling initiatives. The Certified AI Engineer designation offered by ARTIBA is considered among the few best AI certifications globally, accepted around 186 countries. It assesses candidates on their capabilities in AI technologies and provides them with free learning resources to help them get certified for a prized long-term career.


Another such certification goes by the name of Certified AI Expert offered by Global Tech Council, whose exams are taken in online mode only. eCornell also offers a certification program in Machine Learning, used for systems run on AI. The professional education wing of MIT also provides short-term courses and certification programs in Machine Learning and Artificial Intelligence, which can be taken up in-person for sixteen days on MIT’s campus.


The goal of making AI systems more intelligent falls on its professionals. Are you ready to get skilled and certified to take these obstacles heads on?



I just find myself happy with the simple things. Appreciating the blessings God gave me.

Notice: ob_end_flush(): Failed to send buffer of zlib output compression (0) in /home/timebusinessnews/public_html/wp-includes/functions.php on line 5420