The standout story was Google’s AlphaGo computer program beating world champion Lee Se-dol in the DeepMind Challenge Match at the ancient Chinese game of Go. However, AI’s potential impact on search also piqued the interest of many.
Prior to the Go challenge, most commentators thought AlphaGo might take a leg, or possibly two, so when it trounced Lee Se-dol 4–1, a lot of people were taken by surprise.
One of the talking points was the seemingly unorthodox and aggressive moves AlphaGo pulled early in the games that surprised its human opponent. In many ways it highlights this form of AI’s potential to help human decision making because of its enormous ability to learn in ways unfamiliar to the human mind.
Beyond this exciting challenge, the development of AI in search continues. Google’s RankBrain – the machine learning component of its algorithm – has shown evidence of improving search results. A recent (and excellent) study by Stone Temple Consulting showed that RankBrain has significantly improved Google’s ability to show accurate answers for questions it previously struggled to understand.
This follows the leak to Bloomberg at the end of last year that RankBrain was the third most important ranking factor. All of which has succeed in getting the search community excited about what RankBrain is and how it’s working.
Anyone who checks where their website traffic is coming from is likely to see that non-paid traffic plays a significant part in getting traffic to their site, so knowing how to improve this channel will form a key part of a company’s marketing strategy. If RankBrain is as effective as it would appear, getting to grips with what it is looking for – at a guess, high-quality content – and ensuring a website is set up to deliver this, will be important if Google increases its influence over the algorithm.
However, we wouldn’t discount links still being important just yet. The hyperlinked nature of the web – and Google still relying on link crawling to be able to deliver relevant results – probably means they have a significant role to play.
With many updates due to roll off Google’s production line in coming months, it will be interesting to see how search, and its incorporation with AI and machine learning, develops.