Jaspreet’s Expertise:
Founder of TechWhisperer, specializing in advisory and consulting on digital transformation and technologies like AI, blockchain, Web3, and the future of work. Leading expert in digital transformation, blockchain, the future of work, and AI. Former Chief Digital Officer at Mahindra Group and Regional Director at Microsoft.

The Title of His Talk:

“Can AI be Creative?”

The Provocation:
Jaspreet challenges the common assumption that AI lacks creativity.

Jaspreet’s Background:
Author of the best-selling book “The Tech Whisperer” on digital transformation and enabling technologies. Author of “The Immune Organization” on future-proofing organizations against disruptions like COVID-19.

Provocative Question:
Jaspreet Bindra presents a provocative question: “Can AI be creative as human beings?” Audience members participate in a poll, with a surprising 60% believing that AI can be creative and 40% believing it cannot.

Types of Creativity:
Margaret Bowden’s work on creativity is acknowledged, defining three types: Exploratory creativity: Typically understood as creativity, involving exploration and edge options. Combinational creativity: Taking elements from different fields to create something new. Transformational creativity: Creating entirely new concepts and ideas not based on existing knowledge or experience.

AI’s Potential for Creativity:
Bindra argues that AI has the potential for exploratory and combinational creativity but may struggle with transformational creativity. Exploratory creativity: AI can generate various options and identify edge cases, replicating this type of creativity. Combinational creativity: AI can combine elements from different fields, as seen in Zaha Hadid’s architectural work, which was inspired by natural curves. Transformational creativity: AI’s ability to engage in this highest level of creativity is still uncertain and remains a topic of ongoing research and debate.

AI’s Role in Creativity:
Bindra emphasizes that AI is not meant to replace human creativity but rather to augment it. AI can assist humans in creative tasks by providing suggestions, generating ideas, and identifying patterns. The collaboration between humans and AI can lead to innovative and groundbreaking outcomes.

Conclusion:
Bindra concludes by acknowledging that while AI has made significant strides in creative fields, it is still limited in its ability to achieve transformational creativity. The potential for AI to be creative remains a topic of ongoing discussion and research, with experts exploring ways to bridge the gap between AI’s capabilities and human creativity.

AI Defined:
AI is not a single concept but encompasses a range of technologies and approaches. John McCarthy, the creator of AI, famously said that “as soon as it works, no one calls it AI anymore.”

Challenges in Defining AI:
Matt Velloso’s humorous definition of AI: “In R it’s probably analytics, in Python it’s machine learning, and in PowerPoint it’s AI.” Many claims are made in AI that lack substantial evidence.

Kate Crawford’s Perspective:
Kate Crawford, a writer, professor, and teacher in AI, argues that AI is neither artificial nor intelligent.

Questions about AI Creativity:
Can AI be creative, and if so, will its creativity be similar to human creativity or entirely different? Will AI creativity complement or antagonize human creativity? Will it work with us or against us?

Creativity and AI:
Creativity is often seen as the opposite of AI’s current focus on machine learning, reinforcement learning, and deep learning. Al-Faifai, a famous Chinese artist, defines creativity as an innate human trait that can only be untaught, not taught.

Reconciling Opposites:
How can we reconcile the idea of AI creativity, which is taught through data sets and patterns, with the notion of creativity as something that can only be untaught? Creativity is often associated with childhood and the ability to see the world in new and unconventional ways.

Creative Expression by Humans:
Humans express creativity in many ways, including creating music, painting, playing games, speaking, and writing.

AI Generated Music:
AI can generate music that sounds like it was composed by a human. One example is an AI-generated song that was trained on the Beatles’ repertoire. The song has senseless lyrics but sounds like a Beatles song.

Future of AI Music:
AI-generated music is becoming more sophisticated and can be found on various platforms. There is even an AI generator that creates music for artists who have passed away, such as Nirvana and Kurt Cobain.

Music:
AI is creating high-quality music that rivals human compositions.

Art:
AI-generated art, such as the “lost Rembrandt” created by Microsoft Research and partners, has fooled art experts. The website nextrembra.com showcases the process of creating the AI-generated Rembrandt. AI is generating diverse art styles, including those of famous artists like Salvador Dali. AI’s art capabilities are advancing rapidly, with Meta recently releasing a tool for creating AI-generated videos.

Gaming:
AI’s impact on gaming is well-known, with AI defeating humans in games like chess and Jeopardy. While some may argue that AI’s success in these games is more about computing power than true AI, it still demonstrates AI’s potential in gaming.

AI’s Mastery of Gaming:
AI has achieved remarkable success in playing games, particularly Atari Breakout and the complex game of Go. A breakthrough moment occurred when DeepMind’s AI discovered an innovative “edge technique” in Atari Breakout, demonstrating self-learning and strategic thinking. In the game of Go, AlphaGo from DeepMind made history by defeating the world champion Lee Sedol in 2016. AlphaGo’s “move 37” in the third game astonished experts and highlighted the AI’s ability to make unconventional yet effective moves. Since then, AI engines have consistently outperformed human players in games like Go and chess.

AI’s Proficiency in Natural Language Processing:
IBM’s Watson made significant strides in natural language processing, engaging in a debate with a human world champion debater in 2019. Watson’s ability to understand and respond to complex arguments in real-time showcased its advanced language comprehension skills. The debate highlighted the potential of AI to engage in sophisticated conversations and make persuasive arguments. However, the human debater ultimately won the debate, suggesting that AI still has room for improvement in certain aspects of natural language processing. The use of a human jury to determine the winner raises questions about the objectivity of the evaluation process, particularly when considering the potential for bias.

Generative Writing by GPT-3:
GPT-3, a powerful generative engine, has demonstrated impressive capabilities in text generation, including engaging conversations and responses to complex prompts.

Debate with Human Philosophers:
GPT-3 engaged in a debate with nine human philosophers, displaying a sophisticated understanding of philosophical concepts and engaging in cogent arguments.

GPT-3’s Self-Awareness:
In one exchange, GPT-3 displayed self-awareness and introspection, acknowledging its own strengths and limitations while emphasizing the value it places on certain qualities that humans might lack.

Chapter Written by GPT-3:
The author, Jaspreet Bindra, included a chapter in his book, The Tech Whisperer, that was entirely written by GPT-3, highlighting the potential of AI in creative writing.

Poll Results:
A poll conducted during the presentation showed a shift in audience perception, with more people believing that AI will be more creative than humans.

Questions from the Audience:
The author invited questions from the audience, expressing his willingness to answer those he could and acknowledging the importance of open dialogue.

Ethical AI:
AI’s immense power requires careful consideration and ethical development. Governments and tech giants are publishing guidelines for creating ethical AI. AI ethics and society courses are becoming essential to navigate the complexities of AI use.

Nuclear Energy Comparison:
Jaspreet Bindra compares AI’s potential impact to nuclear energy, highlighting the need for responsible management. Both AI and nuclear energy demand careful handling and regulation to ensure positive outcomes.

AI Ethics and Society:
AI ethics is a pressing concern due to the widespread use of AI for profit motives. Shifting the focus towards societal benefits is crucial for responsible AI implementation.

Domains Unconquerable by AI:
The question of which domains AI will never conquer remains unanswered. Kai-Fu Lee’s book discusses 10 domains that AI may struggle to penetrate.

Apprehension and Encouragement:
Concerns about AI’s future are valid, but the ongoing discussions and initiatives for ethical AI are encouraging. Striking a balance between apprehension and optimism is essential for shaping AI’s trajectory.

Compassion and Creativity Matrix:
Jaspreet Bindra introduced a matrix with creativity on one axis and compassion/love on the other. High compassion, high creativity domains include adult care, nursing, management, and potentially CEO roles. The premise is that teaching AI to love or emote is extremely difficult, if not impossible.

Blind Debates:
Siddharth raised a question about blind debates, but the answer was unknown. Jaspreet suggested it could be like an advanced Turing test, where participants try to determine if they’re interacting with a human or an AI without visual cues. Turing tests have been debated as a reliable method for evaluating AI capabilities.

GPT Blind Tests:
Jaspreet mentioned blind tests using text generators like GPT to determine if the text was generated by a human or an AI. He hadn’t conducted an in-depth study on the outcomes of these tests. He expects mixed results, with some cases where AI can deceive and others where it can’t.

More Questions than Answers:
Preeti noted that discussions like this often leave more questions than answers, which is a sign of a vibrant field. She expressed gratitude to Jaspreet for taking the time to join the discussion despite the unusual hours.