Mailbag: Students’ Questions on AI

Mailbag: Students’ Questions on AI

On the evening of Friday, October 5, UIB’s Toby Ruckert, Google’s Zhao Lilyang, and Firemark Lab’s Claudio Caballero sat down at Google’s APAC headquarters in Singapore to talk about Artificial Intelligence (AI) and answer students’ questions on AI in front of a packed audience. The students asked more questions than Toby, Lilyang, and Claudio could answer, so we wanted to share out UIB’s answers to the questions that couldn’t be answered live.

Students’ questions on AI


How do I get started learning about AI?

Read everything you can. One of the easiest ways to get started is to follow the AI industry’s leaders on LinkedIn to understand where they see the industry going, the challenges they are facing, and the solutions they are deploying.


Is Elon Musk right about AI?

Smart people have always warned us about the unique risks and unexpected consequences of new technologies, including the steam engine, electricity, and more recently, Snapchat. The first step in solving a problem is to clearly articulate what the problems are (and will be) and Elon (and many others) are doing that right now for AI.


How is AI used in the Financial Services industry?

AI is currently being used by credit card companies to both target customers and identify fraudulent purchases on those customers’ accounts.


How will AI impact workers?

Remember the steam engine? New technologies displace workers, but they also create more new jobs than they destroy (if in different areas). While no one knows exactly what will happen or when there is an expectation that we will likely see greater displacement with AI than we have with recent disruptive technologies. The best advice is to cultivate your resiliency and flexibility. Over the course of their careers, people will not only work for multiple companies but will likely also have many different types of jobs, including many that don’t even exist today!  



What’s your definition of AI?

AI is the branch of computer science focused on creating “smart” machines with the ability to imitate intelligent human behavior.


How can undergraduates get started in AI?

There isn’t a major global company that doesn’t have people working on AI and the number of AI startups is in the thousands. ALL of them need talented people with all levels of education.


What’s the future of unsupervised learning?

Supervised learning is like classroom teaching in a closed environment. Unsupervised learning is like the experiences we get and the inferences our brains make out of it. AI will begin with supervised learning and there are many different models. In the future, we will focus more on unsupervised learning. We’ll make some mistakes as we develop a correct feedback model, but that will help us to improve our unsupervised learning.


What’s the difference between AI and machine learning?

With AI, you teach the machine. With machine learning, the machine teaches itself.


What about machine learning and deep learning?

With machine learning, the machine teaches itself how to solve a specific problem requiring thought. With deep learning, the machine teaches itself how to solve any problem requiring thought.


How do you measure the feasibility of AI projects?

Businesses measure the feasibility of all projects by their ROI, AI projects are no different.


Can chatbots replace existing communications channels?

Yes, right now UIB’s UnificationEngine®-powered chatbots are replacing hotels’ printed guides, service providers’ call centers, and manufacturers’ mobile apps.


What jobs will be “AI-proof”?

If you’re a human, your job — whether you’re an artist, a soldier, a teacher, a politician, a banker, a scientist, or an evil villain — is going to be impacted by AI. While not every job will be replaced by AI, in virtually every job you will be working with AI.


What makes for a good chatbot?

Like all good IT projects, a good chatbot achieves its desired objectives, whether that is reducing costs, increasing revenues, and/or delighting users.


How does AI make the really hard ethical decisions?

Here’s the secret. It doesn’t. AI’s ethics come from its program/coding, so AI’s “really hard” ethical decisions are actually the ethical decisions made by its developers.


How can AI be used to solve social problems?

AI is a new tool that can help us to better articulate and solve the world’s most difficult natural (and man-made) problems. This can take the form of both giving us access to new information and helping us to better understand and take action on both existing and new information.



How do you handle AI bias?

The best way to handle AI bias is to create AI teams with a mix of people on them with different backgrounds and experiences. See UIB CEO Toby Ruckert’s article on achieving gender equality through AI.


How does AI work in chatbots?

For UIB’s UnificationEngine-powered chatbots, AI drives the natural language processing (NLP).


What’s a practical use of AI right now that can benefit a company?

An AI-powered, co-working assistant is a practical use of AI that can benefit a company right now (see more practical uses of AI here).


How is data preparation and data cleansing done?

Data cleansing, the first step in the overall data preparation process, is the process of analyzing, identifying, and correcting [messy] raw data. When analyzing organizational data to make strategic decisions, you first start with a thorough data cleansing process. There are now a large number of tools which help us do this.


Thank you for answering students’ questions on AI — one final question, what’s AI’s main challenge?

Right now, AI is pretty stupid. We still need to invest the time up-front to teach it what we want it to know. The moment we can figure out why some people have more “common sense” than others, will be the moment we crack the “intelligence” part of Artificial Intelligence!

About the Author:

Ken Herron
UIB Chief Marketing Officer Ken writes about the latest IoT and AI global news, trends, and best practices.