5 things humans can still do better than machines

5 things humans can still do better than machines

While AI may be on the rise, there are still several fields in which machines are no match for their human competitors – here are five examples.

Is humankind paving the way for its own obsolescence? This question, which would have sounded outright bizarre – or at least highly speculative – just a few years ago, is increasingly being asked within academic think-tanks and high-profile meetings around the world. The latest wave of automation is having a dramatic impact on the labour market, the consequences of which are only predicted to grow over the next years. It’s not just the taxi-drivers, entry-level white-collar workers and precarious digital employees who are supposed to fear for their workplace: even highly specialized activities seem ever more at risk of being automated, and machines are starting to outclass humans in fields that used to be strictly reserved for them.

Nevertheless, most researchers remain optimistic, cautioning journalists and politicians against premature dystopian scenarios and mounting fears of a robot-apocalypse. Closer inspection shows that generalized artificial intelligence may still be several decades away from becoming a reality; and in the meantime, even basic questions, like the exact nature of intelligence, are still subject to complex philosophical debates.

We took a closer look at five fields that demonstrate the specific qualities of human intelligence, showing how some activities may just be too complex for our fellow machine companions to imitate anytime soon:

Reading: Despite early sensationalistic claims to the contrary, machines can’t actually read better than humans. While they did perform very well in a dataset-based comprehension test, the results only account for a very narrow set of conditions which were specifically designed to give machines an advantage over humans.

Translation: As any occasional Google Translate user can confirm, humans are still the default choice when you need to translate important documents. AI-driven translations, while constantly improving, can lead to unintentionally hilarious outcomes, so it may not be the wisest idea to rely on them when organizing a high-profile summit – as the fiasco of Tencent’s translation system at the recent Boao Forum clearly showed.

Human Resources: Like many other office jobs, human resource management involves numerous activities that are perfectly suited for automation (for instance, handling and archiving large quantities of data). Freeing HR-professionals from this tedious aspect of their work could allow them to dedicate more time and attention to cultivate skills that are peculiarly human, such as empathy and interpersonal relationships – fields which, for the time being, clearly remain a human prerogative.

Art & Music: The idea of creativity as a semi-miraculous, Promethean and quintessentially human quality is deeply rooted in our collective subconscious, so we shouldn’t be surprised that attempts by AI-programs to create original artworks have been met with a mixture of amusement, scorn and unease. Questions concerning the criteria for appreciating AI-generated art, or how the very concept of art could be affected by eventual AI-breakthroughs in this field, remain unanswered as yet.

Emotional Intelligence: Unsurprisingly, the rapid advancement in AI-related technologies is being matched by an equally rapid rise in the fortunes of concepts like empathy and emotional intelligence – the ability to recognize feelings (both in oneself and in others). Despite appeals for the development of “computers with empathy”, emotions remain a deeply human characteristic, unlikely to be mastered by machines in the near future.

Getty Images
Machines still can’t read better than humans (Photo: Getty Images)