To AI or not to AI? My pending question at this year's Learning Technologies at Excel had only timid responses...
Report by Fabrizio Cardinali, LEA Scientific Advisory Board Member.
After nearly 30 years spent on the vendor and researcher side of Learning Technologies, today more than ever I believe that AI is the next big thing which will revolutionise Education making Machines and Humans learn from each other, generating a value adding spiral of knowledge transfer and cross augmentation between human and artificial intelligence making each others life easier.
During the early days of the EU RD program (who remembers the great days of ITS, Intelligent Tutoring Systems calls in the 2nd EU RD Program, Esprit?) we developed tens of Adaptive and Expert Systems for Education, showcasing that the automatic and engaging personalisation of Education could have become possible one day. But in those days both devices and algos where too heavy and clumsy to make it into real school backpacks or corporate cockpits, let alone to the arms of students in the form of smart wearable watches or to the eyes of trainees in the form of lightweight goggles with foldable visual guides and performance aids..
In summary my (educational) dreams were left in my (RD) drawers and folders, but I was sure that the availability of low cost wearables and powerful AI chips would have made it to the market sooner or later, and that Education would have been one of the first fields to embrace AI and IOT, because no other sector would have deserved automated intelligence more than Education to advance the availability of smart machines for end users, let aside perhaps only medicine and healthcare.
Today, with the light speed convergence of Machine Learning, AI and Neural Nets into edge computing hardware and wearables we are witnessing in other fields (e.g. mainly Industry 4 and Healthcare) I was sure I could re-open the drawers and folders of my first love, Learning Technologies, and see flocks of AI providers and vendors issuing new educational appliances offering unprecedent levels of learning intelligence and smart personalization to world students and workers....
With this heavy load of hope I was visiting this year's Learning Technologies 2019, now back in its original venue at London Excel sure to find tons of AI in the booths. I was there as an invited correspondent by the EU Initiative LEA (Learning Exchange Accelerator (https://www.learntechaccelerator.org/), an interesting new support action of the Commission grouping public procurers, advanced suppliers and learn tech experts unite to speed up public procurement of "smart" LT for learning personalization in the Educational sector. The project will be issuing public procurement of innovation (PPI) tenders for personal learning systems in schools in the next future, hence I felt as having the right "business case" to excite many "smart providers" and "first movers" to convince them to disclose me their AI plans with no fear or backstepping (btw...to be kept updated on LEA next calls for proposals you may enroll at the LEA newsletter at https://www.learntechaccelerator.org/followlea/).
Hence, with my "AI for all" hope in mind, I started attending the many free seminars and demo pitches vendors were delivering in the event arena, sure to find evidence that new "exponential" technologies such as AI, ML, blockchain and IOT had finally entered the market of LT, following the trend evident in other key industrial and societal sectors.
Actually my expectation proved to be very unattended: in the nearly 200 free seminars and 30 demos held by exhibitors in the 2 day event, only 7 seminars explicitly had "AI" in the title and only 1 demo was listing AI features upfront. Only one provider was present in both seminar and demo area waving AI messaging in its marketing (the Italian SaaS LMS provider, Docebo LMS, https://www.docebo.com, whose speaker Josh Squire, Director of Enterprise Solutions, https://www.linkedin.com/in/joshsquires opened the seminar with an intriguing title , "Give a warm welcome to your newest team member, Artificial Intelligence", listing the possible pillars of AI in Education...).
Few other vendors presented listing AI explicitly in the title, introducing mainly two, quite "traditional", uses of Machine Learning outcomes in Education, namely :
- Course Suggestion Engines (e.g. one presented by Open Sesame, the US provider of elearning content courses, https://www.opensesame.com and another one by Hownow, http://gethownow.com/, both showing similar suggestion capabilities to the ones we use in our daily quest for preferred content in B2C markets such as Netflix, Youtube & Spotify amongst the others).
- Training Assistive Chatbots (presented by a bunch of vendros , all more or less based on Google's Tensorflow and/or IBM's Watson existing NLP engines with very limited versioning for educational and teaching skills, despite their nicknames)
A commendation should be given to SABA, the renown long time LMS vendor, who turned its seminar in a tips and hints session on Neuroscience, enlightening good teaching strategies and clues on how trainees learn better whilst adapting to different teaching stimuli.
A pity the Saba workshop missed the link to Artificial Neural Nets (ANNs) and how they may be used to learn automatically what stimuli are more successfull in a personal learning experience. And even more my hope of finding a clear leadership for AI in Education vanished when the speakers gazed at me after I timidly asked if they were planning to embrace such interesting corpus of neuroscience knowledge in some sort of AI powered assistant or alike for future Saba LMS users...
In summary, apart some timid examples, AI seemed to me being the big missing name at this year’s LT 2019 event in London. Even if, wondering around the seminar backstges and stands layouts, many if not all vendors were declaring they had AI plans and pilots which surely will make it to public stands and announcements next year, few vendors knew what they were talking about when I start asking them which type of learning approach, deep or shallow, they were planning to train their learning machines and how they were able to provide better human teaching, truely. Let alone their inability to explan on which data set they were training their smart machines and with which scientific evidence of effectiveness.
In conclusion, my "hope tour" finished with a strong feeling that if AI has the track record to be the next big thing in Education, it had yet to land in the world of Learning Technologies providers. Will we see it at next year's LT event? I think so, since if there is one thing that smart humans and machines ought to share amongst themselves, this is intelligent learning practices and know how.
I can wait to see....but perhaps the market should accelerate adoption: Humans ought to design new revolutionary learning techniques and tools that AI surely makes feasible, before machines learn how to embrace them alone …and then it will be too late!
Find this article also on LinkedIn.