Challenges arise from power of artificial intelligence
After the sweeping defeat of the world’s best Go player by a computer program, and with science fiction depictions of a future in which robots supplant their creators seemingly not so far-fetched, perhaps it is time to think about setting boundaries
Fears of an artificial intelligence takeover have been raised afresh by the AlphaGo computer program’s sweeping three-game defeat of the world’s best Go player, China’s Ke Jie. It caused ripples last year by beating another master, South Korean Lee Se-dol, and has been secretly trouncing competitors in on-line competitions under a pseudonym. Although its creators have now retired it from the game, troubling questions have been raised: Where next will machines claim a victory against humans and will the successes end? With science fiction depictions of a future in which robots supplant their creators seemingly not so far-fetched, perhaps it is time to think about setting boundaries.
AlphaGo’s success lay in its ability to teach itself how to improve. Ke was adamant before the games that he could win, but did not count on the program learning so quickly from its mistakes. The victory was given even greater meaning by being inflicted during a five-day summit titled “The Future of Go”, in the Chinese town of Wuzhen, the birthplace of the fiendishly difficult game. But perhaps there should have not been too much surprise; from Google’s search engine to on-line personal assistants like Cortana and Siri to self-driving cars, AI is progressing rapidly.
Ke believed letting emotions get in the way of his game was his biggest mistake. But DeepMind, the Google-owned company behind AlphaGo, has achieved what it wanted and is not interested in a rematch. Attention will now turn to using the program to help scientists tackle challenges like finding cures for diseases and improving energy efficiency. Those are comforting thoughts that detract from the reality that the more humanity embraces AI, the more challenges that will arise.
Most at risk are jobs, although while AI-based technology will eliminate some, it will create others. But there are related issues like security, privacy and ethics that have to be considered. As interactions with intelligent machines increase, what values they should be embedded with have to be given careful thought. Super-smart devices are still some way in the future, but given the complexities, we need to start thinking about them now.