An Entrepreneur, Coach, IT Consultant, Strategic Adviser, and a Traveler craving to explore and contribute to forming a better society.

Sunday, July 30, 2017

Artificial Intelligence: Hype vs Reality and a few questions

No comments :
Artificial Intelligence is making news every day. Numerous online medias have made it really difficult to separate the reality from hypes and myths. I came across the slides from Luminary LabsStarting with the most common myth "AI is coming to take your jobs", these slides from Luminary Labs has discussed a variety of myths about AI and tried to explain. It's an interesting read. Before you read this post further, I suggest you, to go through the slides.

Image Source: slideshare.net

Employment

Should you go deeper into the content, it raises a lot of questions that are not answered. For instance, these slides claim that "AI will augment human" and it also declares "Human jobs won't go away, but it will change. Roles will be more creative and specialized". This naturally raises a series of questions: what will happen to people who are not creative? What if you don't have any special skills?

These slides claim that "the more machine we employ, the more people skills we need". Really? If yes, how? Probably we may need few people to manage the machines, but what happens to the rest? 

Automation

The airplanes have already adopted the automation during flight. However, the cars on road pose a completely different challenge. Some algorithms forces at least one should be in the driver seat to control in case of a malfunction. During a lunch with one of my client, an executive officer representing from a reputed organisation, raised an important question: "Imagine that you are writing an AI program that deals with the driving algorithm, what would the logic be in a situation where the car has to hit a human to avoid an accident?". Should the car save the passenger? or should the car save the other human being on the road? 

One may argue that in a fully automated world, this situation may not arise. My question to them: what should happen while we are in the beginning stages?

Recommendation algorithm (Machine Learning)

Machine learning definitely has its benefits without a doubt. For instance, the promises to read human health data for better prediction is definitely a boon to doctors and society. ML is already helping us a lot in finding friends in social networking site and finding your next interesting movies and the techniques employed by Netflix like companies are quite effective. I liked the majority of the movies suggested by the algorithm. 

Sometimes, it goes overboard by presuming that you will like Bollywood cinemas if you are an Indian. This is the algorithms' failure to understand the cultural diversity of a country like India. I believe this can be corrected as the algorithm evolves. How to teach the machine to understand cultural diversity and regional intrinsic values of a person?

A student that I met shared his opinion: "You can only teach the machine logics, but you can't teach it to be creative." What is your opinion? I believe it is too early to predict AI's impact, but we can only imagine now.

No comments :