Bulgaria’s best potential ally isn’t the EU. It isn’t the US or Russia. It isn’t China or the UN. It isn’t mountains or pasture, rivers or seas, sun or stars. Bulgaria’s best potential ally is artificially intelligent (AI) life.
AI life is what machines become when they stop doing only what we humans tell them to do and start thinking and doing in ways we humans can hardly imagine. We’re seeing glimpses of it already. A Google-created computer network recently taught itself the game of Go, the most complicated game known to man, just by watching the game being played and experimenting with strategies it made up. It taught itself so well it beat the world’s best human master of Go 4-1 in a five-match series.
Ok, I hear the objections. Go is just a game. Beating us at it doesn’t liven up an electronic number cruncher any more than flying in airplanes turns passengers into angels. The more computers do silly or mundane thinking for us, the freer we humans are to indulge in higher creative pursuits like the arts. If you want to give machines truly human challenges, ask them to paint like Picasso, van Gogh, Kandinsky, Turner or Munch! That will separate truly higher intelligence from the AI wanna-bees.
Last year, three researchers at the University of Tubingen did just that. They introduced AI beings to some famous paintings, encouraged them to identify distinctive styles, and asked them to apply the styles to a photograph of the Neckarfront in Tubingen. Here are the results:
In the big picture, we ain’t seen nothing yet. The driverless cars that drive us safely, the AI investors who look after our finances, the AI physicians who look after our health – they’re not ready for mass use. Within a generation they will be.
What will our world look like then? It’s hard to imagine. So it’s easy to flip from imagining AI life won’t happen – it will! – to fearing it will run amok. OMG, 50,000 dead ringers for Arnold Schwarznegger in The Terminator! Like Luddites clubbing factory looms, we’re tormented to kill the new before it grows.
Yet AI life will grow, we’ll come to appreciate it, and we’ll no more long for its disappearance than we’ll long to make all our clothes by hand. And if, heaven forbid, 50,000 Terminators march on us, our best defense will be 50,000 more advanced De-Terminators we have enlisted to defend us.
In fact, our most pressing problem with AI life is that we’re not tapping it quickly enough. That’s not AI’s fault; it’s ours. Human society desperately needs more workers skilled in machine learning, Big Data, and data science. If you don’t know what those phrases mean, you’re not alone – five years ago hardly anyone did. For quick definition, machine learning is the core of artificial intelligence – helping computers learn for themselves rather than just feeding them instructions. Big Data refers to data sets billions of rows long and thousands of columns wide, far too big to fit into any single computer processor’s memory. Data science, once better known as informatics, harnesses big data to make useful predictions.
Most cutting-edge advances in AI life are currently emerging from the US, particularly from Silicon Valley. Silicon Valley boasts the highest concentration of talent, the densest network of information sharing and the richest backing by informed venture capitalist. Yet US schooling in general – the top echelons excepted — is so mediocre, so bureaucratic, and so obsessed with race and gender that domestic supply of AI-competent works falls way short of demand. Moreover, the huge profits generated by high-tech firms and shared with top workers drive local housing prices sky-high; employees earning 100,000 a year feel second-class.
In Europe, London is the main hub of AI innovation. Like Silicon Valley, London meets a huge share of employer demand by importing skilled workers, mainly from elsewhere in the EU. Between lesser support from venture capitalists and London’s lure for skilled Europeans, AI-related pay and benefits tend to be lower in London than in Silicon Valley. Yet London’s preeminence in global finance keeps housing prices high and squeezes real purchasing power.
Imagine the opportunities for Bulgaria! Bulgaria has a strong math and computer science culture, a well-developed IT infrastructure, valuable experience subcontracting with global IT firms, and a good reputation for software development. It has many English-language speakers — for although AI life doesn’t care, English is currently essential for good communication with US- and UK-based AI firms. Bulgaria offers cheap housing, low costs of living wand modern amenities. It enjoys free trade and relatively free movement of people within the EU and access to EU development funds. Tax rates are low, legal contracts are generally enforced, and major violations can be appealed to the EU.
Are those ingredients enough to make Bulgaria a haven for AI development? No. Like other countries, most of its current programmers need retraining to handle Big Data, data science or machine learning. A new generation needs to be trained. Teachers themselves need retraining. That can’t happen overnight.
More important, opportunities never force themselves into anyone’s life. People have to seize them and rarely do. Why not? Habit, inertia, familiarity. If a new approach is so good, we wonder, why isn’t everyone else already doing it? Mother Nature noticed this shortcoming long ago, and devised two clever ways to deal with it. One is the cycle of life and death: the replacement of older generations with younger. The other is hunger: an intense desire for something we don’t have.
How about it, Bulgaria? Is being the poorest country in the EU good enough for you? Do you want to let other countries and cultures bear the risks of AI development and wait for benefits to trickle down to you? Or do you want to embrace new AI life, actively promote its development, and savor the joy and material rewards of discovery? In short, would you rather ban driverless cars or help design, test and build them?
The choice is yours.
By Kent Osband