Rise of the Machines: Surprising things AI is already doing19 February 2019
Beyond Siri, Alexa, and scary walking robots, AI is moving into all areas of consumers’ lives. Here are five of the more unusual applications we found. 1. Making beer Creating alcoholic beverages might not be the first thing that springs to mind when you consider artificial intelligence. But making beer is what IntelligentX Brewing Company’s systems are doing. They put the customer at the heart of the manufacturing process, connecting their feedback to an AI brewer. Recipes are adjusted based on the input, creating an improved, personalised beer. Many applications of AI focus on using huge datasets to automate processes. IntelligentX’s beer offer instead uses AI to enhance products for individual customers and create a meaningful connection with the brand. 2. Cutting hospital waiting times AI is well known for its ability to improve diagnostic results. But now that ability is being harnessed in a new way to improve the dreaded hospital waiting list. NLP algorithms and the latest image recognition abilities of AI systems were used by WMG at Warwick University to train a system to recognise abnormalities in chest X-Rays, prioritise those positive results and then recommend a response time from a consultant. 3. Turning human brains into a ‘swarm intelligence’ Much hype is made of the potential of AI to replace human intelligence. UnanimousAI has taken a different approach. Their system uses AI to connect many human brains into what they call a ‘swarm intelligence’. Early applications of these ‘swarms’ have been encouraging with some very accurate predictions made – here’s their founder describing a famous example which must have had the bookies spitting. Human experts connected as a system in this way, can make more accurate predictions than they would as individuals, and also more accurate than if they were to reach consensus as a group. 4. Discovering a new species of human The existence of a third human ancestor other than Denisovians and their more famous cousins, Neanderthals, was only a theory. No longer. By combining deep learning algorithms with statistical methods and applying them to the human genome, researchers have identified the footprint of a new hominid in the genome of Asian individuals. Their find was backed up by the discovery of a fossil in summer 2018 that fits the characteristics of the theoretical ancestor. 5. Teaching children about emotion Computer vision and sound processing capabilities are now at a stage where they can complete extremely complex assignments, such as deciphering human emotion. Affectiva’s system uses deep learning to do just that. One of the interesting uses this has been put to is PeppyPals, a game which teaches pre-schoolers social skills they can use in the real world. It teaches them about EQ without using any text or language and can be used to help diagnose conditions such as autism and ADHD.