Showing posts with label Artificial Intelligence (AI). Show all posts
Showing posts with label Artificial Intelligence (AI). Show all posts

Monday, February 19, 2024

Case Study on promise and perils of AI: Imran Khan gives AI generated ‘Victory Speech’ from Jail

 Not a day goes by without a headline grabbing story about Artificial Intelligence (AI) and its innovative application. Not surprisingly, I was fascinated to “see” the “AI version” of former Pakistani Prime Minister Imran Khan’s victory speech posted on Youtube after his party won an unprecedented election. What was remarkable was the innovative use of technology that led to this win.

Mr. Khan who is currently serving a sentence on charges of corruption holed up in a Pakistani prison wasn’t allowed to contest elections or even engage directly in any political communications. But in a groundbreaking move that marked a paradigm shift in political communication, he continued to rally his troops via AI morphed images and synthesized speeches crafted with the assistance of cutting-edge artificial intelligence tools. In a grand finale, he took to the “virtual” stage, thanking his audience in an “AI version” of what he termed “an unprecedented fightback from the nation.”

This is indeed a fascinating application of AI/ML in politics out in the open – a topic at the intersection of Technology, Business, Government and Society. The mainstream media around the world and political analysts took note. The New York Times describes the victory speech as “the mellow, slightly robotic voice said in the minute-long video, which used historical images and footage of Mr. Khan and bore a disclaimer about its A.I. origins.”

Innovative use of AI tools and technologies

Using sophisticated deep learning models, AI can now analyze and synthesize audio and visual content with remarkable accuracy which are hard to distinguish from the real-version with a naked eye. This capability enables the creation of videos that seamlessly blend real and virtual elements, blurring the lines between fiction and reality. From realistic computer-generated characters to immersive virtual environments, AI-generated videos push the boundaries of what was once deemed impossible in visual storytelling.

Hollywood has long used AI-generated videos in streamlining the film making process, especially to generate special-effects that wow us. With the ability to automate certain tasks, such as scene composition, special effects, and even scriptwriting, AI can significantly reduce production timelines and costs. Filmmakers and content creators can leverage these technologies to bring their creative visions to life more efficiently, fostering a new era of innovation in the entertainment industry. The technology emerged from Hollywood to social media and is now into the mainstream influencing mindshare in politics and democracies.

Influencing mindshare


Imran Khan’s use of AI-generated videos to broadcast political speeches are at the intersection of innovation and society. These are not merely a string of algorithmically generated sentences but a reflection of the leader’s vision, values, and commitment to progress in Khan’s ‘voice’. The AI system, trained on vast datasets encompassing Khan’s speeches, policy documents, and public addresses, successfully captured the essence of his distinctive communication style.

Critics and supporters alike lauded the innovation, acknowledging the efficiency and effectiveness of AI-generated speeches. While skeptics questioned the authenticity and spontaneity of the delivery, supporters emphasized the potential for AI to enhance and amplify the impact of political messaging.

While ethicists are lauding the disclaimer that came with Khan’s speeches, one wonders about all DeepFakes circulating without a similar notice. What if Mr. Khan’s team had used the same technology to generate DeepFakes of his political opponent, Nawaz Sharif, in a compromising position with Bollowood starlet or worse making blasphemous speeches?

The Ethics


The widespread use of AI-generated videos, especially to influence large groups of people raises ethical considerations and challenges. The potential for deepfake technology, where AI can convincingly manipulate footage to depict events that never occurred or alter the words and actions of individuals, introduces concerns about misinformation and trust in the digital age.

As these technologies advance, concerns regarding misinformation, privacy invasion, and the potential misuse of AI-generated content have emerged, prompting a critical examination of the ethical implications surrounding this cutting-edge field. Striking a balance between the creative potential of AI-generated videos and the responsible use of this technology becomes crucial in navigating the ethical landscape.

While the technologies to detect deepfakes continue to advance at a fast pace, bad actors bank on the virality of social media. It is almost like a game of real-life whack-a-mole where the good-guys try to catch-up and swat deepfakes sprouting around. The bad guys who create deepfakes bank on their viral spread on social platforms. Attempts to later contradict them with facts backed by digital proof of tampering can’t undo the widespread damage done.

Of course, this is not the last word on the topic. We are sure to see armchair-critics of politics continue to debate these applications as citizens of two other major Democracies – the largest and the oldest – go to polls this year. The innovative and ‘ethical’ use of the tools by Imran Khan’s team during the Pakistani elections are being studied by political wonks and consultants around the world.


Originally published in Express computers - The role of Artificial Intelligence in Influencing Democracies  

Friday, December 9, 2022

I decided to use an Artificial Intelligence (AI) software for a promo-video for my new book. Here's' the result

Artificial Intelligence (AI) technologies and tools have been advancing at a fast pace. Many new tools continue to evolve and they are increasingly targeted at improving productivity of end consumers. 

After my book - Diary of a Successful Loser - was published last month, I had been reaching out to reviewers and bloggers. Along the way, I decided to explore the use of of Text-to-Video AI software to generate a promotional video for the book. 

Here is one version of a resulting video : What do you think?




While the resulting video is not bad, it is not as slick as one from a creative team or professional videographers. On the flip side, it is much better if not similar to what I would have got by paying a gig worker on Fivrr or Upwork 

After researching for a bit, I decided to use the freeware version Steve AI an online Video making software that creates Videos and animations in seconds. The UI is intuitive and easy to use after a quick signup. 

The challenge with Text-to-AI software is the human element and creativity - the creator (you and I) need a robust yet simple script with the right keywords that we can feed into the AI software. One can choose from a variety of templates, video and voice formats. After that, it is a matter of playing around with different formats, voiceovers and narratives. 

Bottomline: Image synthesis has great implications on creative arts and creation of visual art similar to what smartphones did to still camera. I came away impressed with the ease-of-use of such Ai software, and how they can be a powerful aid to less-creative people. With software like With steve AI’s video maker one can quickly create Facebook ads, video slideshows, newsfeed videos, stories, and cover videos. 


There are a number of 'Text to Video' AI software with varying levels of usability including 

  • OpenAI's DALL-e AI - Open AI announced that it removed the waitlist for its DALL-E AI image generator service. More than 1.5M users are now actively creating over 2M images a day with DALL·E—from artists and creative directors to authors and architects—with over 100K users sharing their creations and feedback in our Discord community.
  • Make-A-Video - Meta's Make-A-Video is an AI-powered video generator that can create novel video content from text or image prompts, similar to existing image synthesis tools like DALL-E and Stable Diffusion. Make-A-Video research builds on the recent progress made in text-to-image generation technology built to enable text-to-video generation. The system uses images with descriptions to learn what the world looks like and how it is often described. Link to Meta's research paper 
  • Stable Diffusion - A newly released open source image synthesis model called Stable Diffusion allows anyone with a PC and a decent GPU to conjure up almost any visual reality they can imagine. It can imitate virtually any visual style, and if you feed it a descriptive phrase, the results appear on your screen like magic.
  • Imagen Video - Google’s newest AI generator  that creates HD video from text prompts. Google's engineers claim it is a text-conditional video generation system based on a cascade of video diffusion models. Given a text prompt, Imagen Video generates high definition videos using a base video generation model and a sequence of interleaved spatial and temporal video super-resolution models. According to Google's research paper, Imagen Video includes several notable stylistic abilities, such as generating videos based on the work of famous painters (the paintings of Vincent van Gogh, for example), generating 3D rotating objects while preserving object structure, and rendering text in a variety of animation styles. Google is hopeful that general-purpose video synthesis models can "significantly decrease the difficulty of high-quality content generation."