Not AI
Everyone’s selling AI. Or are they?
The learning conference merry-go-round kicks off again this year with Learning Technologies next week. No doubt many of the usual taglines will be unfurled, along with a few more this year about AI. AI-driven, AI-enhanced, AI-powered, AI-ultra-turbo-super-powered...
But not all that glitters is gold! Here are some testing questions you might like to put to the AI-merchants as they descend on you in exhibition halls around the world this year.
What advantage does it give, really? Examples of AI beating human intelligence are still rare (chess, go, reading comprehension) which is partly why there’s so much hype when these breakthroughs do happen. So it’s quite possible that it just adds a layer of black-boxed complexity to your business and actually destroys rather than adds value. You need to be convinced that the cost-benefit analysis comes out positive vs manual, vs basic automation and vs do-nothing. Can the AI-merchant demonstrate how the AI performs against these?
Does it get better over time? Machine learning is just a machine getting better at performing a task with experience, over time. So AI-merchants selling machine learning should be able to answer the questions: does the AI get better over time?; by how much?; And at what rate? A great AI-merchant might be able to relate the answer directly to data consumption.
Is it really all that clever? A simple if statement stuck at the top of a digital product (are you busy? yes/no) can be truthfully described as an adaptive technology. It would be a stretch to call that AI but some merchants do. What does the AI comprise of? How many algorithms are there? What purpose does each serve?
Does the AI-merchant own the AI? There are a lot of free, sophisticated AI and ML tools out there now. It’s not a problem, necessarily, if your AI-merchant use some of these. But if they are, there’s a limit to the competitive advantage you can get from technologies that everyone else can freely access (unless you make especially smart use of it). You should also be able to negotiate robustly with re-purposers of open-source. If the AI-merchant has a patent, there’s a greater chance that the tech is indeed both original and sophisticated. Patent-awarded means a lot more than patent-pending. Anyone can apply for a patent, thereby achieving pending status for a number of years.
AI is such an overused term it now has little meaning. You can get the meaning back through dialogue though and I hope some of the above helps with that.
See some of you next week.
===
Put Filtered to the test! Our Chief Scientific Officer, Chris Littlewood, has 15 years experience in data science and will be delivering a session on surrounding AI, personalization and chatbots. Pose questions and visit Stand D11 to get a free demonstration of patented AI at work.
===
Find other articles like this
This post is part of a series that my colleagues and I at Filtered are working on. It’s broadly about how recommendations help us to make sense of all the content clutter, especially in learning. Have a look at our other posts here.
New L&D community
We’ve recently launched a free new community for L&D professionals on Slack. Every Tuesday we’ll be holding live discussions at 4 PM GMT on topics including virtual classrooms, content curation, artificial intelligence and more.
Intelligent learning recommendations
Last year we launched a new tool to provide personalized learning experiences for L&D and HR professionals called globalfilter for L&D. It's an online recommendation engine with over 200 high-quality learning assets to read, watch, practice and apply for our industry. Click on the link above to try it out, or trial our version with 500+ assets for all professions here.