Mahesh Khatri
Aug 31, 2019 · 5 min read

Tweet thoughts on implementing Artificial Intelligence (AI)

Image for post
Image for post
August 15th 2019 Mumbai sunset photograph captured by author with panorama automatically created by Google’s AI Assistant in the author’s phone. It also retrieved the photograph using the search word ‘sunset’.

On the evening of India’s 72nd independence day on 15th August 2019, I participated in a live Twitter (#MITSMRChat) chat on ‘Implementing Artificial Intelligence (AI)’ organized by MIT’S Sloan Management Review. Sharing below my answers to some questions asked in the chat discussion.

What is the definition of AI ?

There is no standard definition of AI. However, earlier this month, the US based National Institute of Standards & Technology (NIST) released a document titled ‘U.S. LEADERSHIP IN AI: A Plan for Federal Engagement in Developing Technical Standards and Related Tools’. The following AI definition is mentioned there :

While definitions of AI vary, for purposes of this plan AI technologies and systems are considered to comprise software and/or hardware that can learn to solve complex problems, make predictions or undertake tasks that require human-like sensing (such as vision, speech, and touch), perception, cognition, planning, learning, communication, or physical action. Examples are wide-ranging and expanding rapidly. They include, but are not limited to, AI assistants, computer vision systems, biomedical research, unmanned vehicle systems, advanced game-playing software, & facial recognition systems & application of AI in both Information Technology (IT) & Operational Technology (OT).

Is AI just hype ?

AI has moved much beyond hype. Too many developments are happening both at scale & across scope. Consider this metric as per my earlier Medium article. For a simple human task such as object detection in images, Stanford Dawn December 2018 benchmarks show that AI is over 1000 times faster & 100 times cheaper than a US worker. Also, experts agree that a lot of AI research work is happening behind closed doors of say Chinese or other state institutions or global giants. Many AI milestones being achieved by organizations are possibly not being published publicly. So, AI has moved beyond hype, for sure.

Can the US & India collaborate on AI ?

Yes. The above NIST document’s Executive Summary (Page 6) does state the following : “Accelerate the exchange of information between Federal officials and counterparts in like-minded countries through partnering on development of AI standards and related tools”. Globally, India is uniquely positioned with strengths both of China (strong national technology savvy leadership) & the US (a large number of democratic institutions). Even though India lags behind both the US & China now, we can catch up with intelligent AI implementation by leveraging our huge IT talent base & the fast growing per capita mobile data consumption.

Will China beat the US in AI ?

US still has a huge lead in AI innovation. The US (like India) has many democratic institutions encouraging innovation & bottoms up thinking which is crucial to address the ongoing jobs & societal disruption. Also, experts such as Vinod Khosla believe that US will innovate in next generation small data AI solutions better than China.

Our world is already being disrupted today. What is AI’s impact ?

With AI, the pace of disruption becomes even more exponential. New AI leveraged capabilities helps propel organizations into functional competencies & domains not ventured before. Past experience is no longer an advantage compared to smart Mathematics leveraged scale enabled AI functionalities. Many organizations have not yet understood the size & scale of the ‘Unknown Unknowns’ macro competitive dynamics AI factors at play.

How does my organization create an AI strategy ? Which areas should be of focus ?

As Professor Ajay Agrawal of University of Toronto’s Rotman School of Management says, AI is about predictions getting better & cheaper. By leveraging data. Both structured and unstructured. Start small, but with smart measurable goals. Choose a strategy which is simple to articulate & gets a buy in from all organizational levels & a visible path to any or all of the three broad organizational goals of either Increasing Incomes, Cutting Costs or Satisfying Stakeholders. Also, it would be perhaps easier to start with tasks first rather than processes.

The greatest impediment to AI implementation is that business people do not fully understand AI’s capabilities & limitations. They do not get to see, touch, and work with AI.

The mobile phone is the simplest teacher to help learn AI. Many examples of daily used phone applications like Google Maps or Gmail or Google Photos & others are getting easier to use day by day due to AI. As an illustrative example, all photographs on this page were fetched from my mobile phone using simple English search queries without any tagging on my part.

Image for post
Image for post
Image returned by Google’s AI Assistant when queried for the term ‘Software’ in the author’s mobile phone

Google’s AI Assistant automatically learnt to figure out the content in both the photographs shown above and below. It saved me time.

Image for post
Image for post
Google Photos displays this picture when queried for ‘Singapore sunset’ on the author’s phone.

The key for business people is to understand that AI is just a tool to help them. AI is just Augmented Intelligence’.

What is a key challenge in implementing AI systems ?

A key challenge in real world AI production systems is the ongoing efforts to refresh the model architecture and parameters based on ongoing model monitoring efforts. Real time & “near” real time data poses challenges both in terms of the end user absorption perspective & also in terms of AI infrastructure scaling.

The field of AI is fast changing. How do I as a software technologist keep pace with the same ?

Ability to learn / unlearn new coding / programming languages & tools is equally crucial to survive the tectonic AI technology shifts. Already, a lot of automation is happening within the AI field itself. The long term key to survive and grow is to grasp generic knowledge irrespective of specific technologies & tools across the software & AI landscape.

Can AI be used for societal good ?

Yes. Hopefully it will happen. AI is too powerful a technology not to be utilized for common good. Obviously, the dangers of powerful enemy state actors or global giants misusing cannot be overlooked. Capitalist laws of markets will not change with AI. Just as it has happened with Open Source Software markets, there are & will be many sources of free or extremely low cost AI solutions that will be available for use by motivated individuals for societal good.

I welcome your thoughts and comments on this article.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch

Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore

Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store