Chatbots and the Human Element

I have just started a new section of Northeastern’s EDU-6323: Technology as a Medium for Learning.  This is the accelerated 8-week version for summer, covering an array of digital technologies and their integration into K12, higher ed, or corporate training sessions.

During the first week, the assigned readings included:

Hubert’s post garnered the most comments in the ensuing discussions in class.  Hubert suggested that chatbots and AI could impact education through:

  • Automatic Essay Scoring
  • Spaced Interval Learning
  • Conversational Student (and Student-Centered) Feedback
  • AI Teaching Assistants (Ms. Watson)
  • Chatbot Campus Genie

My students’ reactions were quite bipolar.  Some were excited and saw this as positive…others were dismayed and saw this as negative.

On the positive side, one said that the idea of a virtual assistant was exciting.  Another noted that most student evaluations of courses are poorly written, and that a chatbot might draw better assessment data out of students.

The positives were outweighed in the discussions by negative thoughts.  One student saw the use of chatbots as de-personalizing and de-humanizing education.  Another noted that in medical education, much of the essay grading involves partial credit…and she questioned whether an algorithm could capture the nuances a human professional brings to grading.  One questioned whether the introduction of digital assistants might undermine the role currently filled by graduate assistants and post-docs.  One stated that adding personal “genies” to students undermined those students learning critical life skills…such as getting from point A to point B without a smartphone!

Good points, but let me add to the positive side of the conversation.  There have been a number of blog posts recently on similar themes. The first by “Emma Identity” – a bot – discusses how big data can determine any individual’s writing style after 5000 words…which in essence would make it impossible for anyone to plagiarize.  AI would know whether a student had or had not written something.  A definite positive!

A more detailed look at the potential of AI was by Lucas Rizzotto in The Future of Education.  A very long piece, but in essence it suggests that AI could create a sweet spot between personalized learning, mastery learning, and experiential learning.

In Why Do Chatbots Give Us Hope for the Future?, the author noted that chatbots are always on, accessible, transparent, and logical.  Couple this with one of Mary Meeker’s 2017 slides that I noted in my last post:

 

As I noted in the last post, I found the year-to-year growth of voice queries mind-blowing…and again, it raises questions for me about learning management systems, learning activities, and how … to channel Richard Mayer … we might tap in to dual-channel learning.  It does suggest that we are already globally comfortable with chatbots!

The post Virtual Assistants and What You Can Do With Them differentiates between chatbots and virtual assistants like Alexa, Siri, or Google Assistant.  The author suggested that chatbots are more narrowly defined apps than virtual assistants.

“Chatbots are beginning to get a lot smarter, but for businesses, their primary function is as a virtual agent for a specific app, brand, or service. Chatbots help customers do things such as book travel, shop and complete e-commerce transactions, or get customer support information and submit helpdesk tickets through a conversational interface. If a chatbot is a virtual agent set to task within a specific app, then a virtual assistant is what happens when you give the AI free reign throughout an OS.”

Two books that I have read recently around smart technology are Martin Ford’s (2015) Rise of the Robots: Technology and the Threat of a Jobless Future, and Kevin Kelly’s (2016) The Inevitable: Understanding the 12 technological forces that will shape our future.   One scary…one hopeful…and both directly applicable to this course!  But both are clear on one point – the world is changing!

To close, one student made the point that the world is inflexible and students need to get used to it.  Given the rise of artificial intelligence, I would counter that because of the increasingly interconnected humans and things, the world is actually becoming quite flexible…and we need to get used to it.

But…I am biased.  What are your thoughts?

{Graphic: Maciej Lipiec}

Defining Online: Ask the Machines?

Dave Weinberger had a very interesting post on Backchannel last week that suggest AI now has knowledge we will never understand.  Dave noted:

“We are increasingly relying on machines that derive conclusions from models that they themselves created, models that are often beyond human comprehension, models that “think” about the world differently than we do.”

He goes on to say that we have long been trying to simplify the world – think Einstein attempting to find a Unified Field Theory to tie together relativity, electromagnetism and gravity.  This new machine way of thinking may suggest “simple” is wrong.  Google’s AlphaGo program can now beat anyone playing Go…but cannot teach you how to play Go.  It thinks differently than its human competitors.

Dave noted that for thousands of years, we acted as if the simplicity of our models reflected the simplicity of our universe.  We are beginning to learn that with an almost infinite number of interrelated variables, the real world is too complex to “know.”

In Tom Friedman’s book, Thank You For Being Late: An Optimist’s Guide to Thriving in the Age of Accelerations, he suggested that Mother Nature provides a good model for facing the future, as Mother Nature has been adaptable and resilient for billions of years.  Mother Nature has shown the ability to adapt when new knowledge becomes available, the ability to embrace diversity, the ability to balance micro- and macro-level processes, and the ability to change in a sometimes brutal fashion.  In other words, Mother Nature crunches the numbers and does not attempt to simplify the models.

I am not sure why..but this all came to mind as I read Tony Bates’ post “What is online learning: Seeking definition.”  Tony described a new survey going out to all Canadian institutions of higher education, seeking to understand the future direction of elearning in Canada.  He noted that simply defining “online learning” has proved problematic, as different institutions have somewhat arbitrary definitions of online, blended, hybrid, and even the differentiation between credit, contract, and free courses.  Over the past two decades, I have run into similar issues trying to define what is meant by online learning.  I have run the gamut from structured courses run asynchronously (and sometimes synchronously) through LMSs…to “It’s All Frigging Online!” – meaning that we are all now so interconnected that “online” is simply a continuum by which learning can be facilitated…but that continuum rarely approaches zero.

Tony noted that the results will be available in early September, and I suspect his team will learn from these results.  I also wonder if sufficient data already exists in the cloud that machine intelligence could look at the same issues and come to new understandings which might be difficult for us to understand?

It is amazing that we live in an era where contemplating that is even possible!

{Graphic: The Vital Edge}

An Accelerating Future


Over the past couple of weeks, I have been exploring the 2017 Deloitte Global Human Capital Trends report, which looked at the challenges ahead for businesses and HR professionals.  The report is based on analysis of a survey of more than 10,400 business and HR leaders globally, and noted ten trends.  Over a series of posts, I have been looking at this report from a faculty development perspective, but folding in thoughts generated from reading Tom Friedman’s new book, Thank You For Being Late: An Optimist’s Guide to Thriving in the Age of Accelerations.

The last two trends involved Diversity/Inclusion and the Future of Work, which again tie in nicely to the accelerating technological changes of the past decade noted in Friedman’s book.  Taking them in reverse order, the report noted that the future of work is being driven by the acceleration of connectivity and cognitive technology.  Friedman noted that between 2000 and 2007, we had a phase shift where “…connectivity was fast, free, easy for you and ubiquitous and handling complexity became fast, free, easy for you and invisible.”

Think about that statement.  When the world went flat, all corners of the world were connected and connected with high bandwidth.  At the same time…and due in large part to those connections and the continued advances of Moore’s Law, machine learning made the potential for handling complexity effortless.  With the huge data available on the cloud, Google can now translate English into any language (and vice versa)…not by programming grammar…but by letting the program compare examples of translated text and look for statistical patterns.  Friedman pointed out that when Google got rid of linguists and brought in statisticians, accuracy of language translation went up.  Now the same thing is happening with speech recognition, and we are approaching the old Star Trek standard of a universal translator.

This use of the cloud has allowed for some amazing transformations in what is “normal.”  Friedman quoted Tom Goodwin in a TechCrunch article as stating:

“Uber, the world’s largest taxi company, owns no vehicles.  Facebook, the world’s most popular media owner, creates no content.  Alibaba, the most valuable retailer, has not inventory.  And Airbnb, the world’s largest accommodation provider, owns no real estate.  Something interesting is happening.”

Now, conceptualize how that “something interesting” will play out in higher education.

The Deloitte report noted that automation, cognitive computing, and crowdsourcing are paradigm-shifting forces that will reshape the workforce.  With AI impacting almost every field, every field will have to identify those “essential human skills” that will differentiate their business and make them competitive.  This obviously will also impact what higher education is doing to prepare the workforce of the future…which in turn impacts what faculty need to do.  The report suggested that the essentially human parts of work – empathy, communication, persuasion, personal service, problem solving, and strategic decision making – are becoming more important, which raises the importance of a diverse workforce.  The report noted that when one considers organizations as networks, it becomes clear that diversity and inclusion can enhance organizational performance.  And diversity is not just gender or ethnic considerations, but diversity of thought as well.

Now consider faculty development in this accelerating future.

The gold standard regarding faculty used to be tenure-track processes.  But in an accelerating future, tenure is simply a waypoint towards an undefined future.  The half-life of the skills and expertise one brings in to tenure will erode rapidly.  More importantly, thanks to cognitive computing, some aspects of “teaching, research and service” could easily be automated.  This is not bad.  Friedman points out that the future will involve teaming of humans with machines.  Rather than a TA, we might have Siri or Alexa or some other cognitive device to help us … and learn with us.

That suggests that faculty – and faculty developers – should be asking:

  • What parts of teaching, research, and service can be automated, and what parts do faculty provide added value?
  • How do faculty reskill … and help students reskill as technology evolves?
  • What learning needs to take place in a classroom and with students physically present and what could be done online?  Synchronous, asynchronous, small group, simulations…
  • What new learning can be (or should be) crowdsourced?  What does this mean for curriculum design?
  • With all this change, time becomes a precious commodity.  How do we redesign faculty (and student) work to be open, collaborative, digital…and yet leave time for exploration and discovery?
  • Will new roles emerge beyond tenure-track, term, and adjunct faculty?  How will faculty development evolve to meet these new roles?  With the world moving to more personalized experiences, will we now have personalized faculty development?

No easy answers…but complacency could be our biggest barrier.  We have to assume that the faculty development model of the past will not fit an accelerating future.

{Graphics: Deloitte Press}

Attracting FacDev Talent

I have been exploring the 2017 Deloitte Global Human Capital Trends report, which looked at the challenges ahead for businesses and HR professionals, but I have been looking at it from a faculty development perspective.  The report is based on analysis of a survey of more than 10,400 business and HR leaders globally, and noted ten trends.  I discussed the second of these – careers and learning – yesterday.

The third trend involves talent acquisition…which at first glance does not have much to do with centers for teaching and learning and faculty development…or does it?  The report noted that in…

“…today’s transparent digital world, a company’s employment brand must be both highly visible and highly attractive because candidates now find the employer, not the reverse.” (Emphasis mine)

In ten years in faculty development, I have been involved in many search committees for members of CTLs.  I am sure many of you have as well.  The time honored process of crafting and posting a job description, forming a committee, screening a large number of applications…many of which do not fit the requirements, phone and maybe web interviews, campus visits, and the hope that through all of this, a candidate that actually is a good fit will be found.

The Deloitte report suggests this model may be changing … that tech solutions may disrupt this process.  AI systems like IBM’s Watson can now sort through cloud networks like LinkedIn and quickly identify good fits based on career experiences, endorsed skills, and analysis of social media dialogue.  Organizations are already employing simulations and gaming into the interview process to analyze potential performance on the job.  The report noted that “…a consensus is emerging that traditional interviewing – subjective and unstandardized – may be an unreliable method for predicting a potential employee’s success.”

Joel Osteen has been quoted as saying “See, when you drive home today, you’ve got a big windshield on the front of your car. And you’ve got a little bitty rearview mirror. And the reason the windshield is so large and the rearview mirror is so small is because what’s happened in your past is not near as important as what’s in your future.”

Perhaps the way we have staffed CTLs in the past is our rear-view mirror, and the future staffing of CTLs might involve leveraging technology, focusing on the center’s brand to attract new talent to desire to come to the center, and thinking outside the box to find the right talent that can help faculty enhance student learning.  We tend to think that the past is crystal clear and that the future is fuzzy…just like the picture below.  True…but the future is also always not what we expect…so staffing for what we expect seems out of sync.

If you were starting a CTL from scratch now, what are the talents you would want with you as you look to this future?

{Graphics: Deloitte Press, Bill Frymire}

UPDATE:  After hitting publish yesterday, FastCompany published “The War For Talent is Over, And Everyone Lost.”  It took a slightly different tack than I did, but it illustrated that organizations seem better able at waging war on talent as opposed to attracting talent.  This article noted that talent is largely personality in the right place…which brings me back around to the idea of making CTLs attractive and the right talent will find you…as opposed to the other way around.

 

Smart and EdTech

A few days ago, I discussed my upcoming doctoral class on leadership in a wired world.  I will also be teaching a Masters of Education course starting next week.  EDU 6323 at Northeastern University is entitled Technology as a Medium for Learning.  We explore aspects of digital technology through the lens of Michelle Miller’s (2014) book, Minds Online: Teaching effectively with technology.  My course (adapted from one developed by Stan Anamuah-Mensah at VCU) flows like this:

I would like to think that this course explores “smart” uses of digital technology for learning…but “smart” has nuanced meaning now.  We are seeing more and more application of AI in our everyday life.  Just examine the Gartner Hype Cycle for Emerging Technologies.

Smart dust, smart workspace, smart data discovery, smart robots…there are a lot of smart applications emerging!  Two books that I have read recently around smart technology are Martin Ford’s (2015) Rise of the Robots: Technology and the Threat of a Jobless Future, and Kevin Kelly’s (2016) The Inevitable: Understanding the 12 technological forces that will shape our future.   One scary…one hopeful…and both directly applicable to this course!

So I read with interest a short article this week in MIT’s Technology Review by Hossein Derakhshan.  In “A Smarter Web”, Derakhshab suggests that we need more text and links, and fewer images, videos and memes.  He noted how the early days of the web and it’s text-based blogs served to nurture varying opinions, while lately, social media apps like Facebook, Twitter and SnapChat have served to amplify existing beliefs, polarizing and fragmenting society.  He suggests that the lack of varying opinions had more to do with the outcome of the recent USA Presidential elections than false news.  Derakhshan suggests that a smarter web would be one that stepped backwards in time.

As I mentioned in my previous post, we are moving as a nation into a future where the old rules seem to have shriveled.  Robert Reich noted earlier this week that Trump’s tweets are becoming a new form of governing by edict..or “Tweedict” as he termed them.  In my EDU 6323 course, we will be using Twitter as a form of class communication using the hashtag #EDU6323…but hopefully in a more collaborative way than the Tweedict suggests!

In the third week of the course, we will explore validity on the web.  I am adding danah boyd’s recent Medium article – “Did Media Literacy Backfire” – to the reading for that week.  danah noted that media literacy asks people to question information and be wary of what they are receiving…but, in line with Derekhshan’s article, this has led to where we are questioning so much that we talk right by each other.  danah ends her article noting:

“The path forward is hazy. We need to enable people to hear different perspectives and make sense of a very complicated — and in many ways, overwhelming — information landscape. We cannot fall back on standard educational approaches because the societal context has shifted. We also cannot simply assume that information intermediaries can fix the problem for us, whether they be traditional news media or social media. We need to get creative and build the social infrastructure necessary for people to meaningfully and substantively engage across existing structural lines. This won’t be easy or quick, but if we want to address issues like propaganda, hate speech, fake news, and biased content, we need to focus on the underlying issues at play. No simple band-aid will work.”

So I know that my course will engage students and introduce them to new technologies.  But my hope is that my course will also begin to shift some paradigms and shake up some cultural norms.  Otherwise, their future students might not hear these different perspectives that they need to hear!  And that would not be smart!