Chess teams that pair humans with machines beat humans alone and beat unaccompanied machines. The lesson is that workers should not fear being replaced by technology, says.
While Facebook boss Mark Zuckerberg hails artificial intelligence’s “potential to make the world better”, many others are far more cautious.
Tesla boss Elon Musk has warned against machines taking over the world. Reluctance to technical progress is not a new phenomenon.
In 19th century Britain, the Luddites, worried about being replaced by technology, sought to destroy textile machinery.
The fear persists, all the more since AI is improving by the day, including at tasks previously thought of as exclusive to human intelligence.
In 1997, the IBM computer Deep Blue spectacularly defeated world chess champion Garry Kasparov for the first time.
Many commentators were shocked by the prowess of AI in a strategy game often referred to as the ‘king of games’.
AI has moved beyond being a super calculator and most advanced machines will be prominent in strategic thinking and creativity.
While many people worry about being overtakenby some form of Skynet, the artificial intelligence defence system from the Terminator film franchise, the last 20 years of computer usage in the game of chess actually suggests the future could be a lot more positive.
Chess was one of the first areas tackled by AI, making it an interesting case study. In hindsight, AI brought chess and its human players to previously unattainable heights.
Advanced chess computers helped humans improve their own skills. Nowadays, top chess players spend most of their time analysing the game via computers.
AI can be used to re-evaluate positions that were previously misunderstood or to rule out moves that are inefficient and focus, instead, on more promising game plans.
This evolution occurred despite critics lamenting the end of chess when the machines first defeated man in 1997.
Though it was feared at the time, chess players have not given up on the game.
They trained harder and became stronger than players of the past. Rather than competing against AI, chess players utilised it.
The benefit of teaming up with AI is illustrated by a new game mode, cyborg chess.
It is named after the cybernetic organism, an entity with both organic and mechanical body parts.
As in a typical chess match, the human players face each other across a chessboard, but in cyborg chess they also each have a computer, running chess engines.
While AI is superior to the human brain in a one-to-one chess contest, human players still contribute to the team.
Humans may let the machines make most of the calculations, but, ultimately, they have their own understanding of the game.
In some situations, human players make better decisions than machines, and successful cyborg chess players know when they can let the machine decide on the move to play, and when they shouldn’t.
Hence, the best cyborg chess team is higher ranked than the best chess engine. This means that the association of human intelligence and AI outperforms stand-alone AI.
Another crucial teaching of cyborg chess is that there is a specific skill set for collaboration with AI.
The ability to work efficiently with AI matters more for cyborg chess players than their standalone strength in chess. Several cyborg chess competitions were not won by the strongest attending chess players.
For instance, in the PAL/CSS Freestyle Tournament in 2005, three chess grandmasters could notdefeat the team Zacks, comprised of two average club players using less powerful AI than the grandmasters.
However, the latter were extremely well-prepared. They had trained extensively with several chess engines and had selected the one they understood the most.
In a one-to-one game againsta grandmaster, these two amateurs would have less than 1% chance of winning.
Hence, this stunning performance suggests that although mastering chess is a valuable skill for cyborg chess, it is not the main skill.
Instead, fruitful collaboration with AI is the key skill.
This example teaches us that humans and machines have complementary capabilities.
Machines supervised by humans are often capable of doing more than machines or humans on their own.
This is becoming common in many fields: plane pilots assisted with auto-pilot programmes, computer-assisted surgeons for complex surgeries, and many others.
AI will increasingly be used in the near future and automated tasks will bring forth new jobs.
A recent study of the Institute for the Future and Dell Technologies states that 85% of jobs that will be occupied in 2030 do not exist yet; just as countless jobs that exist today were unimaginable 20 years ago.
These new jobs will most certainly revolve around this specific human-machine collaboration skill.
Since machines first defeated humans 20 years ago, the evolution of chess suggests that AI need not be feared; it can be embraced and pave the way for continuous improvement in all fields.
The key is to switch from competition to collaboration with AI. Businesses should neither turn away from AI nor simply replace employees with machines.
Success will come from creating adapted teams of men and machines working together.
We humans should primarily focus on improving our understanding of AI by honing our human-machine collaboration skills.