Market Insight by Frédéric Le Hellard - Artificial intelligence: myths, realities, challenges, limits

From science fiction to science fact, artificial intelligence (AI) is now very much a part of our lives and is no longer just the stuff of fantasy, fascination and fear among the public. According to the International Data Corporation (IDC), AI represented a colossal global market of $350 billion in revenues in 2021, growing at an average of 18.8% per year. Indeed, PWC estimates that AI will add $15.7 trillion to the global economy by 2030. It is also a question of world supremacy between the great powers. The rivalry between China and the United States is often highlighted as a kind of cold war. The reality, however, is quite different: the United States is far ahead in the AI race, and Europe is holding its own as the 3rd largest player. 


Deciphering artificial intelligence


For many of us, AI remains mysterious, where machines may gradually take over from the human brain. In fact, this is a relatively old field of research which first emerged in the 1950s. It was symbolic AI (formal logic and knowledge representation) that first launched the discipline, but recent progress is linked to numerical AI (statistics and mass data manipulation). It is the latter that is on the rise, with 89% of the 55,000 patents filed in AI in 2017 , thanks in particular to the formidable developments in computing capacities and the power to process gigantic volumes. According to François Chollet , head of AI development at Google, “we will soon have trained language models on all humanly available texts”.


To solve a problem, the human brain uses two modes of operation in combination, namely intuition and logical reasoning. Surprisingly, AI tends to excel when it comes to intuition, but is poor at reasoning. Everything that a human does immediately and “without thinking” could be replaced by AI: driving in an ordinary situation, using language, recognising objects. AI has already demonstrated some very impressive in the medical field, for example, being able to detect a cancerous tumour on an X-ray much more efficiently than the human eye .