Interviews

Interviews with AI professionals to provide additional perspective.

--

Click here: AI and the Future of Work: a discussion with Irving Wladawsky-Berger

Discussion in relation to a conference that took place at MIT with various AI professionals and researchers, and the question of the impact of AI on the job market. If you’re interested in the question, “What’s Going to Happen? And When?”, I highly recommend reading the related blog post: http://blog.irvingwb.com/blog/2017/12/ai-and-the-future-of-work.html

--

Jim Spohrer - IBM

James C. Spohrer is a computer scientist best known for having led the development of a new science of service systems, often known as service science, management and engineering. In spring 2017, Spohrer was named as Director, Cognitive OpenTech for IBM. From 2009 through 2016, he had been the Director of IBM Global University Programs Worldwide.

Interview

Q: Jim, what does your typical customer profile look like, if there is one?

All entities that use and/or contribute to "open AI code + data + models" (a type of resource) as part of a service they offer to others.

Q: How long do you think it will be before AI trickles down significantly from enterprise into medium and small business?

It already has, see Domingos "The Master Algorithm" - or just appreciate the fact when you do a Google search, look at a social media post, buy things from Amazon, consider an online recommendation, use an online map to get somewhere, you are benefitting from years of AI research on pattern recognition, machine learning, knowledge representation, and automated reasoning.   See: https://www.amazon.com/Master-Algorithm-Ultimate-Learning-Machine/dp/0465065708

Q: How far do you think the current approaches in neural networks can take us before there is some kind of plateau?

Deep learning is a type of multilayer optimization that requires a lot of data and a lot of compute power.   With enough data and enough compute power,  pattern recognition models can be developed that are as good as individual people (typically 5% error rates) and in some cases super-human performance levels.  The next challenge is episodic memory and commonsense reasoning - do we have enough data? do we have enough compute power? No, not yet.  Are the current algorithms alone sufficient? Unlikely that they are, but there is a lot of investment to try to find out more. We will know more once we achieve petascale on data and algorithms at around $1000 cost - estimated by 2040.  We are still at the terascale level (what $1000 buys),and petascale is 1000x terascale.  The human brain is estimated at the exascale, and that is 1000x petascale.   Exascale is estimated to be $1000 cost by 2060.  For more see: http://service-science.info/archives/4741

Q: What would your advice be to college students heading towards the job market, in terms of choosing a career in light of what AI can accomplish? 

Think about the service you can offer to others in society, based on your interests.  Find others with your same interests, who have been succeeding in offering a service to others and learn as much as you can from them.  Find and learn from many, many role models.  Students should start thinking about making a job and not taking a job.  In order to "think beyond taking a job or making career" requires understanding aspirations or even callings.  First remember AI will make access to knowledge cheaper, and that is good thing for tackling unsolved challenges both technical and social.  If you have a STEM (Science-Technology-Engineering-Math) leaning, learn to work on multidisciplinary teams to tackle unsolved problems.  If you like working with STEM people, but are academically business-oriented, art-and-humanities oriented, or social sciences oriented, find a way to be helpful working on multidisciplinary teams to tackle unsolved problems - yours skills are needed to in many ways, from communications and marketing to making sure organizations of people run smoothly.  At IBM we term, this type of person a T-shaped person, with problem-solving depth, and communication breadth.   The world needs lots more T-shaped people.   

When I mentor students, I ask them to come up with a good answer to the question - "What would you do if you have 100 workers workng for you?"   This prepares them for a future of digital workers based on AI, and how they want to augment themselves to offer service to others.

Q: What do you think would be the best path for someone who wants to break out of the gig economy into a more consistent ongoing full time position?

Same as above. Find others with your same interests, who are succeeding in offering service to others and learn as much as you can from them.  To be the best, learn from the rest.  Find role models, and get to know them.

Q: If a student or mid level career professional is open to reinventing themselves, what would be the skills and courses you would recommend that they take? 

Find role models, and get to know them by helping them solve problems.

Q: In light of changes coming from workforce evolution; what advice would you have for colleges to evolve their curriculum? Ground every student in coding and AI, and develop familiarity with the functional areas of business without specializing? 

Lean into entrepreneurship. Recruit entrepreneurial faculty.  Recruit entrepreneurial students.  Create a diverse startup ecosystem that is a thicker-and-thicker ring around your campus.  See: http://www.therepublic.com/2017/09/17/purdue_pushes_entrepreneurial_initiative/

If students are interested in coding and AI great. However, if they are not teach them to be T-shaped with depth and breath, and able to identify role models with related interests, and to help role models solve problems.

Q: Some people look at the trending and call for a tax on automation; others call for universal basic income, and others think it would be impossible to enforce. What’s your view? Do you think we’re headed in that direction, and do you think there is any justification for it?

Many experiments are best, since one size does not fit all.   Taxing automation is just one way to fund a universal basic income, and there are others.  I suggest people read this book: https://www.amazon.com/Accelerating-TechnOnomic-Medium-ATOM-Upgrade-ebook/dp/B072K3JL29

Q: AutoML seems to raise the stakes for what AI can accomplish - do you think it will be possible for startups to compete with larger players? 

Yes, startups will do fine.   The question I am asked more often is "will large companies survive?"  Yes, both will do fine.   Many good books out there on this, but I enjoyed this one most recently: https://www.amazon.com/Get-off-Grass-Kickstarting-Innovation-ebook/dp/B00GF2MOAQ/

Q: Have any movies or books influenced you in thinking about AI? 

See books above.   I liked the movie Her: https://en.wikipedia.org/wiki/Her_(film)

I also blog about a lot of books - see: http://service-science.info/archives/4416