2X Founder & Chief Decision Scientist | TEDx Speaker | Contributor to MIT SMR & Forbes | B-School Adjunct Prof. | C-Suite Advisor
As long as machine learning continues to be ‘cool and trendy’, it won’t be useful. #Business will continue to suffer due to poor value from AI. What comes to mind when you think of ML/AI in real life? Self-driving cars, emotion detection, or maybe predicting your sales figures? These are signs that ML is still in the experimental zone. It shows our bias to use it in suboptimal applications. What are the examples of AI in simple areas? Clicking pictures with your camera. Simple photo touch-ups to fix subtle issues. Gallery listing of your albums. These are applications we take for granted. Most don't even notice the presence of AI in these. When machine learning is applied to common applications it delivers incremental, yet consistent value. Slowly add it to the entire workflow and it can be transformed. Gartner says that by 2022, 40% of ML model development will be done in products that do not have ML as their primary goal. “Technology becomes truly useful only when it becomes invisible.” Machine Learning/AI will have arrived only when it gets boringly commonplace. What are the ‘boring uses of AI’ you notice today? #data #analytics #ArtificialIntelligence
Spam email classification...one of the earliest implementations of ML
AI is present from the era of Nokia cell-phones. T9 dictionary was an example of the same. Boring use of AI is recommended videos on several platforms on the basis of last watch, one's mood can change and so as the watching preference.
Voice bots (Siri, Alexa, et al) are a truly impressive implementation of NLP in daily life. But they are only replacing one interface with another, much like when icons replaced command lines. I recall witnessing the tipping point for personal computers from hobbyist devices to must-haves -- the key was the arrival of truly transformative applications, at that time the first spreadsheet and word processing tools, which allowed individuals to perform tasks that today we take for granted, but previously had required programming skills and/or highly specialized (i.e. very expensive) equipment. So for me, the real question is whether the set of technologies we often lump together as "AI" will remain limited to simply making interfaces nicer (in addition to voice, and autocorrect as already noted by Sourabh Sharma, there's also autocomplete) or will it allow us to perform tasks that we couldn't accomplish previously.
Auto correct
Navi
4 年Setting up calendar reminders by reading emails from Gmail. For e.g. due date to pay a bill etc.