David Vivancos is author of “Automate or Be Automated”, CEO at MindBigData.com, artificial intelligence & data science entrepreneur, keynote speaker and deep learning algorithmist.
Our talk was on AI – the misconceptions, challenges and our future in a world with implemented artificial intelligence and automation.
1. Please can you give us a little background about yourself and your current role?
I have been creating science & technology companies for the last 25 years, from early internet ventures in 1995 lecturing the first Java course in Spain, to Virtual Reality Endeavors, Electronic Art, Nanotechnology for Space Applications, Education, and many others. Over the last 6 years fully devoted to neuroscience & Artificial Intelligence, advising from Startups to CEOs and boards at Fortune 500s, I am also a keynote speaker with 300+ conferences lectured worldwide, and author of several books, the last one just released “Automate or Be Automated” http://www.vivancos.com/automate
2. What are the most common misconceptions about Artificial Intelligence?
There are many in the two sides of the spectrum from plenty of companies using the “AI” hype word, just for marketing but without any real AI being done, to fears of killer robots that are still in the science fiction realm, but it is true that over the last decade AI has achieved many feats in the human only realm.
3. What are the common uses and applications of AI?
There are also many, form vision related tasks, like object, faces, scenes or even emotion recognition in images and videos, or for self-driving cars or drones to navigate, or natural language processing, to understand our voice or recreate it, or to write news or articles, or recommenders systems to drive sales in the online world, or to help discover new drugs in the healthcare industry for example.
4. What are the trends and challenges that AI and automation face right now?
The trends are to start commoditizing some of the algorithms that have left the research realm and become mainstream, so you can buy or rent them as a service in your cloud provider of choice, also to start including AI embedded in hardware, from smartphones to smart TVs. Automation is a must in the digital realm with the emergence of RPA, and in the physical world with robots as cheap as the human labor cost in many countries, still with limited capabilities but growing fast.
The challenges are related to the lack of enough talent in the field, avoid designing algorithms that are unfair and biased, and also training them with data from all of us, not just with a few, so the data is diverse and representative.
5. What the world of tomorrow looks like?
Once we have more computing capabilities to speed up the research times, for example with mainstream quantum computing, the road to superintelligence will be quite fast. With or without it the world in 15 to 20 years will be probably jobless due to the full automation of the world, this is why it is critical that we all learn and be an active part automating it and not being left behind, this is the essence of why I wrote my last book “Automate or Be Automated” to help everyone to be involved while we still can.
6. How we can protect data and can AI be used as a weapon?
We need to own our data in the first place, and this a difficult challenge nowadays, since monetizing the data we produce is at the core business of the startups that rule the digital world. Besides protecting it, we need to start capturing as much data as we can, so we can use it to automate, as I coined it “You must become a dataset”.
Regarding the use of AI as a weapon, yes it is starting to be used as a tool to discriminate, but since we can have AI in almost any device, some with a cost bellow $10, it will be quite easy to weaponize it, in the physical world and in many settings not just killer drones, its implications are still developing, and the frameworks to contain it almost nonexistent. In the digital world the automation of warfare is a reality and the battle to rule it is already ongoing.
7. Who are the people that inspire you?
There are plenty, but narrowing it down to the A.I field this could be a non-extensive list: Geoffrey Hinton, Yoshua Bengio, Yann Le Cun, Andrew Ng, Ian Goodfellow, Fei‐Fei Li, Demis Hassabis, Jürgen Schmidhuber, Max Welling, Diederik P. Kingma, Ruslan Salakhutdinov, Alex Krizhevsky, Honglak Lee, Pascal Vincent, Hugo Larochelle, Léon Bottou, Nando de Freitas, Samy Bengio, Björn Schuller, Chris Pal, Daphne Koller, Cynthia Breazeal, Oriol Vinyals, Jeremy Howard, Daniela Rus, Jeff Dean, Stuart Russell, Max Tegmark, Rana ElKaliouby, Peter Norvig , Siraj Raval or Tanmay Bakshi to name just a few.