AlphaGo proves victorious but what does this say about the future?

16 Mar 2016 02:19 PM

It has been human versus machine in Seoul, South Korea, over the last week with Google’s DeepMind technology, developed in London, in the shape of AlphaGo taking on Lee Se-dol, the world champion in the Game of Go.

After the fifth and final game AlphaGo has reigned victorious over Se-dol. With three victories under its belt, confirming overall victory, AlphaGo had been looking unstoppable in this tense five-game match of the incredibly complicated game. However the human-race fought back gallantly against the Artificial Intelligence (AI) machine to win the fourth match.

Google admitted that its AI machine made mistakes after being put under pressure by Se-dol. Does this show that human intuition is more powerful than machine learning?

Perhaps not after AlphaGo triumphantly won the fifth and final match to secure the series with a 4-1 victory.

What does this mean for the future of machine learning and the development of AI? AlphaGo got to this stage in its development by studying older games of Go, playing against itself and learning from its mistakes. It would seem that AlphaGo was able to continue along this trajectory and again learn from the mistakes it made during the fourth game and did not repeat them. Google themselves said that the loss in the fourth game was very valuable for AlphaGo. It would appear that this human victory may have in fact made the machine more powerful by highlighting the AI’s machine’s existing weaknesses.

If this pattern of learning continues might we be heading towards a future where AI machines become more intelligent through interactions with humans? Will decisions about the way we work, live, shop and interact with public services be decided autonomously by computers based solely on data?

Current discussions on the role of data in our digital development focus on the rapid increase in the volume of data being created and the privacy and trust issues this raises. Looking ahead the increasing convergence between big data, the Internet of Things and the rise of the Smart City will see the amount of data being collected, processed and managed rise to an unimaginable scale, leading more organisations to use automated data processing and machine learning technologies, to gain insights from the volume of data that will be created. The next stage of our digital development driven by machine learning and AI technologies will therefore raise ethical, legal and possible moral questions that need to be considered about the role of automation and how data is being used to make decisions that will affect all of our lives. While this version of the future might be some way off, it also might be closer than we realise.

On 26 April techUK’s Great Data Frontier conference will be bringing together industry leaders, policy makers and key stakeholders driving the next wave of the UK’s digital development to discuss the changing UK data landscape between now and 2020. This event is an opportunity to exchange views with UK data leaders on the legal, technological and policy issues that need to be considered in order to create a positive regulatory environment that can build a culture of data confidence and unleash a new wave of data-driven innovation and growth. Members interested in attending the Great Data Frontier conference can register here.