When you make historical faces come alive – Times of India

A great spectre haunts the world. Machine-learning algorithms, which facilitate a regime of artificial intelligence enabled machines, have begun governing many aspects of our everyday lives. Machine-learning refers to a series of algorithms, which are ‘trained’ on data to progressively improve in its abilities to fulfill a certain given task. These tasks can range from seemingly innocuous ones like machines which determine what social media post we ought to see first when we log into Facebook or Twitter to the potentially transformative events like the rise of self-driving cars.

From solving global warming related challenges to identifying cancerous cells by inspecting a mammogram — machine-learning programs are progressively transforming entire fields. However, there is also a dark side to this emergent techno-algorithmic complex. Artificial intelligence machines could just as easily be used for aggressive surveillance, ethnic discrimination on the basis of machines that read facial features, and ultimately subvert the life blood of democracies through construction of alternate realities. Like with any new technology, these machine language enabled AI programs are a double-edged sword with potentially dramatic consequences. As Max Tegmark, author and physicist at MIT, says, “I am convinced that AI is either the best thing ever to happen to humanity or the worst thing ever to happen.”

Recently, we saw a brief glimpse of the possibilities of deep learning (a sub-branch of machine learning that extracts layers of key information from raw data) when a genealogy research company ‘MyHeritage’ released a new feature called ‘Deep Nostalgia’, which is an image processing technology licensed from an Israeli company called D-ID. By using deep learning, the ‘Deep Nostalgia’ program manipulates images to render photos into near life-like motion creating an extraordinarily vivid sense of familiarity.

As a personal experiment, I used the program to reanimate images of recent historical figures from India’s past such as Bhagat Singh, Swami Vivekananda, Kasturba Gandhi, Lokmanya Tilak, Sri Aurobindo, and Munshi Premchand, which I posted them on Twitter and went viral. From author and parliamentarian Shashi Tharoor to the actor Suniel Shetty, these images were widely shared or commented on. Since then, people have taken to editing videos, adding Rang de Basanti songs on top of Bhagat Singh and tweeting/writing to me as if I have some special insight into the Indian nation/patriotism and so on. Others have taken to emailing me with photos of their deceased parents and grandparents and asking me to reanimate them. The emotional needs that follow from use of A.I. is a topic we are probably not prepared for.

In due course, it is seemingly inevitable that that film industry as well political discourse will use such technologies to create new narratives — some true, some imaginary and some deep fakes.

The writer is a columnist and author of the newly released novel ‘The Dharma Forest’

A great spectre haunts the world. Machine-learning algorithms, which facilitate a regime of artificial intelligence enabled machines, have begun governing many aspects of our everyday lives. Machine-learning refers to a series of algorithms, which are ‘trained’ on data to progressively improve in its abilities to fulfill a certain given task. These tasks can range from seemingly innocuous ones like machines which determine what social media post we ought to see first when we log into Facebook or Twitter to the potentially transformative events like the rise of self-driving cars.

From solving global warming related challenges to identifying cancerous cells by inspecting a mammogram — machine-learning programs are progressively transforming entire fields. However, there is also a dark side to this emergent techno-algorithmic complex. Artificial intelligence machines could just as easily be used for aggressive surveillance, ethnic discrimination on the basis of machines that read facial features, and ultimately subvert the life blood of democracies through construction of alternate realities. Like with any new technology, these machine language enabled AI programs are a double-edged sword with potentially dramatic consequences. As Max Tegmark, author and physicist at MIT, says, “I am convinced that AI is either the best thing ever to happen to humanity or the worst thing ever to happen.”

Recently, we saw a brief glimpse of the possibilities of deep learning (a sub-branch of machine learning that extracts layers of key information from raw data) when a genealogy research company ‘MyHeritage’ released a new feature called ‘Deep Nostalgia’, which is an image processing technology licensed from an Israeli company called D-ID. By using deep learning, the ‘Deep Nostalgia’ program manipulates images to render photos into near life-like motion creating an extraordinarily vivid sense of familiarity.

As a personal experiment, I used the program to reanimate images of recent historical figures from India’s past such as Bhagat Singh, Swami Vivekananda, Kasturba Gandhi, Lokmanya Tilak, Sri Aurobindo, and Munshi Premchand, which I posted them on Twitter and went viral. From author and parliamentarian Shashi Tharoor to the actor Suniel Shetty, these images were widely shared or commented on. Since then, people have taken to editing videos, adding Rang de Basanti songs on top of Bhagat Singh and tweeting/writing to me as if I have some special insight into the Indian nation/patriotism and so on. Others have taken to emailing me with photos of their deceased parents and grandparents and asking me to reanimate them. The emotional needs that follow from use of A.I. is a topic we are probably not prepared for.

In due course, it is seemingly inevitable that that film industry as well political discourse will use such technologies to create new narratives — some true, some imaginary and some deep fakes.

The writer is a columnist and author of the newly released novel ‘The Dharma Forest’