In 1950 Alan Turing asked a simple question in his paper Computing Machinery and Intelligence, “Can machines think?” The last seven decades of computing, artificial intelligence (AI), and machine learning advancements have responded to that question with a resounding “Yes.”

From IBM’s Deep Blue program to Teslas, Roombas, Siris, Alexas, and now ChatGPT and Bard, the advancements in AI have lived up to and surpassed many of Turing’s early musings. Yet we find ourselves at a pivotal point, wondering if this advancement will fundamentally change how we work, learn, design, write, live, and communicate for better or worse.

Generative AI (GenAI), an AI system capable of generating images, text, audio, video, code, and other media in response to prompts, has turned up the volume on this line of questioning as it challenges what and who we believe can create new things, and how valid and unbiased are the things created.

To understand the challenges and opportunities GenAI offers, it is important to know how we’ve gotten here.

Technology advancements have led us here

GenAI has been enabled by significant advancements in AI technology over recent decades. The first was the development of deep learning, a technique for “learning” inspired by how the brain works via neural networks. At the same time, there were advancements in graphics processing units (GPUs) which allowed for complex tasks to be rapidly broken down into smaller subtasks and continuously performed in tandem. While initially applied to gaming, the calculations used by AI models are very parallel in nature, making them ideal for GPUs.

This combination of deep learning and GPUs ushered in a new AI development and adoption era. Deep learning allowed us to develop new AI models that, in many cases, could surpass human capabilities–image and speech recognition, self-driving cars, fraud detection, virtual assistants, and more. GPUs gave us the computational capacity and scale to create these AI models.

The deep learning architecture continued to advance, which in 2017 led to the invention of transformer-based models that give AI the ability to “remember” by tracking relationships in sequential data like the words in a sentence. Because transformers can “remember” what they’ve “seen,” they can build on this to create new content, not just recognize a face or detect spam. As such, transformers and GPUs have ushered in a new era of AI, the era of Creative AI or GenAI, which is poised to become the next platform shift after personal computing, mobile, and the cloud.

Boundless opportunities

It can be argued that in the same way, the internet brought down the marginal cost of content distribution to $0, GenAI could do the same for content creation. That’s why some estimates show the global GenAI market reaching more than $110 Billion by 2030. Couple that with user demand (ChatGPT reached 1 million users in just five days) and the fact that it’s one of the fastest-growing open-source projects, there is a strong case that GenAI could usher in a sea of disruptive change on par with the internet.

From email copy creation to bug testing, customer support, movie making, legal assistance, and invoice automation, the possibilities of GenAI are unprecedented.

Here are a few recent examples of the transformative nature of this technology:

• Travel: Expedia developed a GenAI travel advisor enabling travelers to ask for recommendations on where to go, where to stay, what to do, and more. So, if you are planning your next family vacation to Maui or looking for tee times at St. Andrews, let GenAI do some of the work for you.
• Shopping: Walmart expects GenAI will “be as big a shift as mobile, in terms of how our customers are going to expect to interact with us.” They have already started to use GenAI in their Text to Shop, allowing customers to add Walmart products to their cart by texting or speaking the names of the items they need as if talking to a human retail assistant.
• Education: At the Khan Lab School in Silicon Valley, a GenAI tutor named Khanmigo helps students move towards finding the right answers themselves.
• Sports Commentator: IBM partnered with The Masters to have a GenAI commentator provide detailed golf narration for more than 20,000 video clips over the course of this year’s tournament.

Who will lead the pack?

We are still in the early days of GenAI, so it’s still being determined who will be the driving force in bringing this technology mainstream and how. While some big tech players like Microsoft, Google, Adobe, and Amazon are already starting to stake their claim, other incumbents and startups are lurking in the wings.

There are a variety of questions about how the competitive landscape will play out, including whether it will be dominated by proprietary models or open source, whether it will foster a whole new set of “GenAI First” applications (think Uber for mobile) or just make today’s applications smarter, and the level of verticalization we might see (or not) across the entire value chain. These all factor into what the competitive landscape might look like.

And given the disruptive potential of the technology, it is also attracting a huge number of new startups that hope to beat out incumbents in tried-and-true ways:

• New market disruption — Go after customer needs not served by incumbents.
• Low-end disruption — Go after customer needs not attractive to incumbents.
• 10x better products — Create products so good incumbents can’t compete.

So how does this apply to GenAI?

• GenAI is different from AI until now. It is 10x better technology, meaning it will enable 10x better startups. We’ve already seen 10x better products from “GenAI First” startups like OpenAI and Jasper.
• Incumbents and startups will battle it out at both the platform and application levels. Attracting the talent needed will be key.

Application startups will be susceptible to co-option by incumbents (e.g., via adding and bundling). They will need to quickly create defensive moats via time-to-market and network effects, focusing on customer needs not served by incumbents and through innovative business models not attractive to incumbents.

Growing pains

While GenAI offers obvious opportunities, it comes with its pitfalls and detractors. There is still much to be known about how AI is trained. OpenAI only says that GPT-4 was pre-trained using both publicly available data (think internet data) and data licensed from third-party providers. The amount of data and where it comes from matters, as there are already copyright challenges, and it will be hard to fully rely on GenAI or, worse yet, deal with misinformation.

Deepfakes abound, from Tom Cruise to the Pope to former U.S. President Barack Obama. A deepfake video can show a politician or celebrity saying anything and be very convincing, as seen in the deepfake video below.

There are also legal hurdles and regulations that will certainly have to be overcome. Just recently, Italy’s data protection authority ordered OpenAI to stop processing local data for its ChatGPT generative AI chatbot. It argued that the company breached the European Union’s General Data Protection Regulation (GDPR) regarding data access and protection of minors. Not to mention intellectual property and ethics concerns.

What does the future hold?

Just as we worked through the land minds of the early days of the internet, I expect we will do the same with GenAI. With the exponential increase in data coupled with compute performance, GenAI is likely on an exponential growth trajectory and will drive an accelerating pace of change in the world around us. As with any exponential change, humans could be better at predicting its future impact. Our brains aren’t used to thinking exponentially, and we tend to extrapolate linearly. As such, we consistently underestimate the impact of exponential technologies. The future will be on us before we know it, and we all need to prepare for it.

Companies need to proactively invest in understanding the technology and how it might impact their markets, customers, products, and operations. Maybe more importantly, how it could disrupt them and where it can be used to disrupt others.

As consumers, we should all invest time learning how to use it and understanding its limitations. It will soon become integral to nearly everything we do.

We shouldn’t be surprised or shocked by what comes next. GenAI will move quickly from the new kid on the block to an entirely new era of human-computer evolution. We will have AI doctors, AI lawyers, AI therapists, AI developers, AI artists and composers, AI actors, AI co-workers, and even AI friends. Some are even predicting this is a precursor to general artificial intelligence and digital lifeforms that will exist and evolve independently of humans, maybe even competing with us for the title of “dominant species” at some point in the future.

And, of course, legal and ethical pundits, governments, and the industry, in general, will need to collaborate closely to ensure the needed safeguards are in place.

Whatever happens, we should also remember that the future hasn’t happened yet, and we all get to create it.