You have already heard all about artificial intelligence and the risks it poses to our jobs in a few years… but have you ever considered the possibilities it could offer us? As these new technologies relieve us of certain tasks, what added value would we need to develop? How should we adapt to remain more relevant than machines in the labour market? And, how can entrepreneurship act as a solution to a future labour market with artificial intelligence?
by: Silicon Luxembourg
photo: Thibault Gilbert
featured: Hoai Thu Nguyen Doan
Listen to article (Part I)
Nyuko is organizing a conference bringing together three experts specialized in their fields of artificial intelligence. The conference will examine the very practical ways in which AI is transforming our world and be an open space for these experts to answer any questions related to AI. No holds barred! Here’s our interview with Hoai Thu Nguyen Doan, Luxembourg Chamber of Commerce’s AI Expert.
What is artificial intelligence? Do you have an exact definition for us?
Artificial intelligence is defined as the set of theories and techniques aimed at designing machines capable of simulating human intelligence. In general, a distinction is made between low AI (technology currently available) and high AI (future technology).
AI is characterized as low if it is a simulation of human cognitive functions. In this context, the AI is programmed to provide a predefined response to a given situation without understanding the context. It, thus, gives the illusion of intelligent behaviour in a specific domain (e.g. Deep Blue or Alpha Go game programs). In contrast, strong AI, which is currently pure fiction, refers to a machine capable of producing intelligent behaviour, experiencing self-awareness, and understanding one’s own reasoning (e.g. R2D2 by Star Wars).
“”Artificial” refers to the result of creation through human hands and abilities, as opposed to what is “natural”.”
Since when have we been using technologies based on artificial intelligence?
The first AI references date back to 8th century BC. In the Iliad, a major epic of Greek culture, Homer describes Hephaestus, the blacksmith god who had designed anthropomorphic gold maidservants (an ancient version of robots) to help him with his work. Then there was also Talos (the “Ancient Terminator”), a bronze giant, responsible for protecting Crete on behalf of King Minos. In 13th century Europe, automatons entertained guests in the great courts. And, in the 18th century, craftsmen raised the bar in virtuosity to create realistic representations of life (ex: Jacques de Vaucanson’s “Digesting Duck”).
AI technology has undergone many developments over time. As historians find traces of automatons built by craftsmen from civilizations all over the world, such as Yan Shi, Heron of Alexandria, Al-Djazari or Wolfgang von Kempelen, these stories validate a fascination with AI very early on.
Is artificial intelligence artificial or mostly human?
“Artificial” refers to the result of creation through human hands and abilities, as opposed to what is “natural”. Creating a child would be considered natural intelligence while creating a machine would fall under artificial intelligence. In both cases, the act is still human.
Listen to article (Part II)
What are the main trends in AI?
The economic, social and societal advances made possible by AI are plenty and as a result, major economic powers around the world are investing significant resources to pursue an AI strategy.
The United States, supported by private investments from major multinationals such as GAFAM and IBM (Google, Amazon, Facebook, Apple, Microsoft, and IBM), holds a position as a world leader in AI research. China, on the other hand, aims to reach a level of investment in AI comparable to that of the United States by 2020 and ultimately become the world leader in the field by 2030. In Japan, AI is one of the cornerstones of support for “Abenomics,”government measures focused on finding solutions to social challenges such as a declining birth rate, an aging population and environmental problems. The European Commission has also allocated funds for AI between 2014 to 2020: 2.6 million euros for areas related to AI (Big data, health, transport, future and emerging technologies, etc); 700 million euros from the European Union and 2.1 billion euros from the private sector for advances in smart robots, 27 billion euros to develop skills around AI and finally, 2.3 billion euros to the development of digital skills.
“Artificial Intelligence has the potential to revolutionize our lives in many ways, from the most innocuous to the most exceptional.”
AI still faces many challenges. If I had to name just a few, I would mention the following:
– The “black box” phenomenon: While the data used to feed the AI and its responses are clearly visible, there is still a grey area about how the data is processed by the machine to achieve its results.
– Lack of “common sense”: Algorithms do not have what we usually call “common sense.” They operate by statistical correlations without checking for a logic. This could generate discriminatory biases. For example, the ProPublica investigation site demonstrated in 2016 that a recidivism prediction tool sold to American courts diagnosed a recidivism rate twice as high for blacks vs. whites. Skin colour has never been part of any official criteria. Nevertheless, due to an accumulation of legacy on historical discrimination, the AI overestimated the probability of recidivism for black populations. Bias is therefore one of the obstacles AI must absolutely overcome.
– Expressing feelings: What we call “feelings” is a hormonal process. For AI to be able to interact with a human being to the point of being able to pass the Turing test, it must still be able to show empathy, and the technology is still in its infancy on this point.
– Data security: An additional challenge for AI is to ensure the security of sensitive data against cyber attacks. According to the 2018 Thales Data Security Report, 94% of companies worldwide store critical data in the cloud and frequently use technologies that are vulnerable to cyber attacks. As a result, there are an increasing number of susceptible areas. This could significantly increase the difficulty of ensuring the confidentiality, integrity, availability and authenticity of data — the fuel of AI.
Listen to article (Part III)
How will artificial intelligence revolutionize our lives?
As AI has many applications, it has the potential to revolutionize our lives in many ways, from the most innocuous to the most exceptional. Whether it is a merchant site that recommends product purchases based on your previous searches or auto-complete on your phone’s texts, AI is already present in various aspects of our daily lives.
Here are some examples:
– In transportation, AI is leading a revolution. According to a study by Oliver Wyman, 20% of our motorized objects today would become autonomous by 2030. The firm expects that there will be more changes in the next ten years than there have been in the last sixty years. AI combined with the Internet of Things and New Energy will be the new way we design cities, urban flows, and autonomous vehicles. These vehicles will enjoy greater energy efficiency, security and quality thanks to predictive maintenance.
“The deployment of AI, far from replacing man with machine as announced by alarmist prophecies, constitutes a real windfall for the labour market.”
– In the health sector, it is already more effective than a doctor at detecting the cancerous elements of melanoma or analyzing MRIs. Fuelled by a wealth of data, AI has the potential to improve medical processes and care in terms of time savings and quality. The exponential computing power of machines will provide access to sophisticated correlation research, allowing physicians to better detect symptoms and predict disease progression. This will enable researchers to gain the ability to anticipate drug side effects as early as the clinical trial stages.
– In the industrial field, AI algorithms are creating the era of Industry 4.0, where production processes are optimized with the the Internet of Things and similar digital technologies (robotics, augmented reality, 3D printing, etc). A cyber-physical production system (in English “CPS”) works as an autonomous system with computer and electronic elements linked to sensors that control physical entities. The German Federal Government has even integrated Industry 4.0 into its “High Tech 2020” strategy.
– In the environmental field, AI would be immensely valuable to combatting climate change and ensuring a transition to more sustainable production patterns. Applications abound today, ranging from regulating the energy consumption of intelligent buildings to protecting biodiversity.
In the short-term, are our jobs at risk?
In 2013, researchers Frey and Osbourne from Oxford University published a techno-pessimistic paper that predicted a threat of automation on 47% of American jobs. A few years later, a study carried out by the OECD concluded that only 14% of jobs were at risk of automation. This study, on the other hand, specifies that 32% of jobs would not be likely to disappear, but would undergo substantial changes in the way they are being carried out. This is a far cry from the catastrophic prediction of robotization.
Of course, some jobs will disappear. Telegraphers, subway punchers or wake-up callers don’t exist anymore, for instance. They’ve been phased out through digital transformation. Rather than job destruction, I’d call it job transformation. In a world characterized by the omnipresence of AI, the nature and content of professions will change.
For example, the March 2017 report of the Californian think tank Institute For the Future (IFTF) estimates that 80% of jobs in 2030 do not yet exist today. The future will probably be collaborative (between man and machine). It will mark the advent of an era in which those who know how to exploit the complementarity between human and artificial intelligence will thrive. The deployment of AI, far from replacing man with machine as announced by alarmist prophecies, constitutes a real windfall for the labour market. This would require action both downstream, by organising technological training for those already in the labour market, and upstream, by revising school curricula in order to train a workforce adapted to the challenges of the future. Preparation and adaptation are key concepts for a successful transition to the labour market of tomorrow.
“What is to be feared is not the reversal of creation against its creator but the use of technology by unscrupulous people to impose influence on others.”
Listen to article (Part IV)
Will my children be led by robots in 2050?
Personally, if I had children, I would be happy to have a full-time “nanny-cooker-chammy robot” at home, like Robbie, the companion robot designed by writer Isaac Asimov in 1940. Unfortunately, we are still far from this stage. No machine has yet passed the Turing test, and strong AI, in 10 years time, will still remain in the realm of fiction.
Will all entrepreneurs be at the service of robots in the near future?
As Albert Einstein pointed out, « Science is a wonderful thing if one does not have to earn one’s living at it. »
I believe that the robot is a tool designed by mankind to serve mankind. What is to be feared is not the reversal of creation against its creator, a.k.a the “Frankenstein myth,” but the use of technology by unscrupulous people to impose influence on others. This is where ethical reflection must take place, followed by the implementation of regulations to protect humanity – the AI army – against itself.