{"id":5621,"date":"2022-10-31T13:56:28","date_gmt":"2022-10-31T03:56:28","guid":{"rendered":"https:\/\/news.sap.com\/australia\/?p=5621"},"modified":"2023-08-17T05:02:42","modified_gmt":"2023-08-16T19:02:42","slug":"ai-in-business-what-you-need-to-know","status":"publish","type":"post","link":"https:\/\/news.sap.com\/australia\/2022\/10\/31\/ai-in-business-what-you-need-to-know\/","title":{"rendered":"AI in Business: What You Need To Know"},"content":{"rendered":"
On the latest episode<\/a> of The Best Run Podcast, I had a chat with Pete Chapman, Asia-Pacific Technology Director and Enterprise Architect for Ernest & Young, and Dr. Kim Oosthuizen, Innovation Principal at the Ecosystem Platform Office for 麻豆原创, about the reality of using artificial intelligence in business environments and what the future of AI looks like.<\/p>\n There\u2019s still much confusion surrounding the fundamental questions of artificial intelligence<\/a>, what is AI, what is capable of and what isn\u2019t it capable of. Many of us were introduced to AI through pop culture depictions such as the friendly robot Rosie from The Jetsons or the rogue servant droid Sonny from I, Robot, which has set unrealistic expectations for their use.<\/p>\n Kim summarises AI in the simplest terms as \u201ccomputational agents that act intelligently.\u201d Performing tasks by using data, algorithms, and programs and processing a specific output or goal. Classified as \u2018narrow\u2019, the AI that exists today can only perform the task in the simple or specific domain its programmed to do so in, it can’t do anything outside of that.<\/p>\n As AI is merely an umbrella term for the vast array of intelligent technologies that exist, we can classify three categories of machine learning AI: \n\u201cWe have a bit of a tendency at the moment to call everything intelligent just because it’s software.\u201d<\/strong><\/p><\/blockquote>\n Passing the Turing Test<\/strong> \u201cThrough a machine learning angle, AI can detect patterns that you wouldn’t normally decipher through traditional statistical algorithms. An example would be, we\u2019ve got a client that was using this technology to process applications for grants. They had a massive backlog that came up unexpectedly, and they used this machine learning to recognise what was tending to get approved and could identify and accelerate those application.<\/p>\n In order to keep things safe and apply their principles, they use the rule that the machine can say yes to something that is beneficial to the customer, but if the outcome is no, then it goes to a human to get double checked and processed. That’s the kind of thing that companies are doing to protect us from the AI getting it wrong and people being disadvantaged by this kind of technology.\u201d<\/p>\n https:\/\/www.youtube.com\/watch?v=SVMG6tOkbF4<\/p>\n <\/p>\n
\n1. Analytical AI:<\/strong> Intelligence that we see 90% of businesses use today, AI that predict, recommend and mine data and learns from past experiences
\n2. Human Inspired AI:<\/strong> Intelligence with some extent of sentiment and reaction that must be programmed, e.g. chat bots
\n3. Humanised AI:<\/strong> Intelligence that understands human emotions and has considerable empathy to respond to the end user in a human-like or natural manner, potentially a technology that will be available very far into the future.<\/p>\n
\nSpeaking to Pete about the real world applications of AI, he made one thing clear, \u201cyou can’t call it AI unless it’s passing the Turing test in some way\u201d. Referring to the test created by Alan Turing to determine whether or not a computer is capable of thinking like a human being. Using this logic as the foundation of our understanding of AI, we can apply this narrow technology to the business world of today.<\/p>\n