5 Eye-Opening Conclusions from my 30-Day Generative AI Journey

“It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of light, it was the season of darkness, it was the spring of hope, it was the winter of despair.”

– Charles Dickens, A Tale of Two Cities

Some people train for marathons, some go on diets. I decided to embark on a 30-day “AI Challenge” instead. The current excitement around generative AI is palpable. I can’t personally remember feeling this energized about something since the 90’s, when I first got my hands on the internet. In my late teens, the web opened new horizons and suddenly the whole world was at my fingertips. It changed quickly too, from just featuring content for consumption to being able to create to your heart’s desire. Generative AI marks a similar shift…much like the evolution of the web from write-only to read-write (and soon read-write-own), it has moved from being concealed under the hood to becoming an accessible and exciting co-pilot for us all.

Since the term was first coined in the late 50’s, AI has been utilized in various applications, such as natural language processing, computer vision, robotics, and decision-making systems. Generative AI uses various machine learning techniques, such as deep neural networks, reinforcement learning or transformers, to learn from large amounts of data and generate new data that is similar in structure and content to the original data. For example, the popular ChatGPT is a generative AI chatbot that can produce natural language responses based on input text, while Midjourney, Dall-E and Stable Diffusion create realistic images based on text descriptions. The reason generative AI is so exciting to mainstream audiences is because it allows us to create new content like text, images, audio, and video using simple natural language prompts, making the tech a lot more accessible.

My own experiments with Generative AI started off with simple novelties like custom horror movie recommendations, reimagining Star Trek characters in hospitality roles, mother’s day poetry, children’s bedtime stories and a rap song interpretation of serious finance literature…but advanced quickly to more creative, ambitious projects (e.g. conceptualizing new products, creating chatbots, dabbling with coding and using ChatGPT as a native language tutor) as my understanding of ‘promptcraft’ grew and I discovered a gamut of new tools.

My month-long odyssey was as illuminating as it was fun. Here are five key conclusions I’ve arrived at:

1. Generative AI is a game-changer for human productivity: Generative AI is the productivity genie we’ve all been dreaming about. It will be omnipresent and deeply integrated into popular consumer software moving forward, superbly collaborative (writer’s block could be a thing of the past) and capable of reducing the amount of time we spend on mundane tasks. We can quickly and cheaply create, summarize, code, monitor, analyze and translate to our heart’s desires. Even in the travel industry, which is typically slow on tech uptake, we’ll see a quick move to integrating AI in improving customer service (think more efficient call centers and actually useful chatbots) to analyzing customer reviews and sentiment, reducing marketing and tech coding costs and even putting better business analytics in everyone’s hands using natural language prompts. Clearly there are going to be a lot of cost and productivity gains that accompany the proliferation of Generative AI…this will have a massive impact on white collar jobs and much like other revolutions, existing jobs will need to adapt, and new roles will emerge. In my own experiments, I found current Generative AI tools to be excellent for drafting and refining content, with plenty of hands-on effort still required to produce desired output.
2. We’ll need to rethink education: Education still feels stuck in a bygone era. Designed for the industrial age and primarily to churn out better workers, we’ll need a serious rethink about how we teach our kids to think critically and leverage new AI tools to augment their learning experience. The educator uproar over potential plagiarism is missing the big picture. Large language models like GPT can finally provide everyone with their very own personalized tutor, be it for learning a new language, computer coding or even having the AI take on the role of a historic figure or philosopher you’re learning about. Combine that with VR and you can bring science, history, and other subject matter to vibrant, fully immersive life. Educators like Khan Academy have already moved quickly to integrate ChatGPT into their platform to create a student tutor and teacher’s assistant.
3. Art will no longer be the domain of just trained Artists: Art has long been considered a uniquely human domain and Generative AI has challenged this long held belief. While many artists are up in arms about how new text to image and video tools are cheapening and plagiarizing their work, there are others who are actively pushing the limits with Generative AI to explore new horizons, from human-AI collaborations (a Neuroscience based AI Opera, anyone?) to more sophisticated, previously impossible multi-modal installations (how about seeing the world through the eyes of an animal?). For ‘left brained people’ like me, Generative AI has allowed me to bridge the gaps in my own specialist talent and training to put my imagination to work and create art, narratives and rich media content that would have been previously impossible.  
4. We’ll boldly go where no man has gone before: AI has incredible applications in science that’ll propel us forward and transform the way we understand and address complex problems. In research, AI is already accelerating scientific discoveries by automating data analysis, identifying patterns, and predicting outcomes. In medicine, AI-powered diagnostic tools offer improved accuracy, enabling earlier detection of diseases and personalized treatment plans. Drug development is also streamlined as AI models predict molecular interactions and analyze vast datasets to identify potential candidates. We’ll see similar advances across multiple disciplines and as a long-time fan of science fiction, it’s incredibly gratifying to see that fiction become reality in our lifetime.
5. There’s a very small, but not impossible, chance this could kill us all: At a bare minimum, generative AI brings up various ethical and legal risks that need robust debate before we can widely adopt and trust these systems. Misinformation through deepfakes and unintentional hallucinations (nonsensical or made-up answers) are one issue. Exacerbating bias due to model inputs are another (after all, these models also rely on the Garbage In, Garbage Out principle). Generative AI can also infringe on privacy and intellectual property by using or revealing sensitive information from its training data set without proper anonymization, encryption, or attribution. Of course, these problems pale in comparison to the threat posed by AGI (artificial general intelligence). Much like the works of Hollywood and our favourite science fiction writers, it is quite possible that we create a superintelligence that results in our own demise, if we’re unable to address the AI alignment problem adequately. Given the high cost and rewards involved, tech giants are already embroiled in an arms race to build the best and biggest model…but newer models like GPT4 have already shown emergent capabilities and ‘sparks of AGI’. This space will also evolve extremely quickly. Nvidia’s CEO recently mused that we’ll see AI capabilities accelerate by a factor of a million in the ten years ahead. Compare this to the pace of regulation and you can see we have a bit of a problem. Much like nuclear regulation, we’ll need strong, comprehensive guardrails and international collaboration, especially given the comparatively lower barriers to entry in this space. We’ll need to move fast to stay ahead of bad actors and hope that the impetus for decisive government action isn’t a (currently avoidable) catastrophe. 

While the last prospect may sound terrifying, overall, I’m upbeat about the potential AI has for us as a species (in my opinion, AI is probably our most important invention since electricity). We’re at a critical tipping point and if we can power through the hurdles ahead, there’s an exciting future that awaits us all. We’ll need to collaborate and adopt a framework for responsible AI innovation, including data governance (quality, diversity and security of data used), model governance (transparent testing and validating for accuracy, reliability, bias and fairness), content governance (ethical and legal use of content, attribution, verifying authenticity and accuracy) and regulatory governance (alignment with law and ethics, protecting rights and enforcing accountability, promoting public awareness and education). As you can see, there’s plenty of work to be done.  

We’re standing at the cusp of a new era. AI is a powerful and promising technology that can help unlock new possibilities for human creativity and innovation. However, much like my Dickens quote of choice at the beginning of this article, the challenges and risks are very real and need to be addressed responsibly. We’ll inevitably hit dizzying highs and nauseating lows. By adopting a framework for responsible innovation though, we can ensure that this technology is largely used for good and not evil, for benefit and not harm, for trust and not distrust.

How do you feel about the current state of AI? Leave a comment or DM me to let me know (I’m always keen to connect with like-minded thinkers and tinkerers). You’ll find all the details of my 30-Day Generative AI Challenge here, with accompanying details on the various ideas, tools and methods used. If you’re so inclined, I invite you to complete your own personal journey and share your progress and learnings using the hashtag #30DayAIChallenge on LinkedIn, Facebook or Twitter.