Join senior executives in San Francisco on July 11-12 and learn how business leaders are getting ahead of the generative AI revolution. Learn more
New products like ChatGPT captivated audiences, but what will be the real lucrative apps? Will they offer sporadic business success stories lost in a sea of noise, or are we at the start of a true paradigm shift? What will it take to develop truly workable AI systems?
To chart the future of AI, we can learn valuable lessons from the radical advancement of the previous technology: the era of Big Data.
2003-2020: the era of Big Data
The rapid adoption and commercialization of the Internet in the late 1990s and early 2000s built and lost fortunes, laid the foundation for business empires, and fueled exponential growth in web traffic. This traffic generated logs, which proved to be an extremely useful record of online actions. We quickly learned that logs help us understand why software crashes and what combination of behaviors leads to desirable actions, like buying a product.
As log files grew exponentially with the rise of the internet, most of us felt we were onto something hugely valuable, and the hype machine went to 11. But there was still see if we could actually analyze that data and turn it into sustainable data. value, especially when the data was spread across many different ecosystems.
Google’s big data success story deserves to be revisited as a symbol of how data made it a trillion-dollar company that transformed the market forever. Google’s search results were still excellent and built trust, but the company couldn’t have continued to provide large-scale search – or all of the additional products we rely on Google for today – until that AdWords enabled monetization. Now we all expect to find exactly what we need in seconds, along with perfect step-by-step instructions, collaborative documents and cloud-based storage.
Countless fortunes have been built on Google’s ability to turn data into compelling products, and many other titans, from a rebooted IBM to the new goliath of Snowflakehave built successful empires by helping organizations capture, manage and optimize data.
What was just confusing chatter at first eventually produced huge financial returns. It is precisely this path that the AI must follow.
2017-2034: the age of AI
Internet users have produced huge volumes of text written in natural language, such as English or Chinese, available as websites, PDFs, blogs, and more. Thanks to big data, it is easy to store and analyze this text, which allows researchers to develop software that can read all this text and learn to write on its own. Fast forward to the arrival of ChatGPT at the end of 2022 and parents calling their children to ask if the machines had finally come to life.
This is a watershed moment in the field of AI, in the history of technology, and perhaps in the history of mankind.
The current AI hype levels are exactly where we were with Big Data. The key question for the industry to answer is: how can AI deliver the sustainable business outcomes essential to drive this sea change for good?
Realizable AI: Let’s put AI to work
To find viable and useful long-term applications, AI platforms must include three essential elements.
- Generative AI models themselves
- Interfaces and business applications that will allow users to interact with the models, which could be a standalone product or a generative back-office process augmented by AI
- A system for ensuring confidence in models, including the ability to continually and cost-effectively monitor a model’s performance and to teach the model so that it can improve its responses
Just as Google brought these elements together to create actionable big data, AI success stories must do the same to create what I call actionable AI.
Let’s take a look at each of these elements and where we are today:
Generative AI models
Generative AI is unique in its savagery, bringing unexpected behavior challenges and requiring continuous education to improve. We cannot fix bugs like we would with traditional procedural software. These models are software built by other software, made up of hundreds of billions of equations that interact in ways we cannot understand. We just don’t know what weights between which neurons need to be set to which values to prevent a chatbot from telling a journalist to divorce his wife.
The only way these role models can improve is through feedback and more opportunities to learn what good behavior looks like. Constant vigilance around the quality of the data and the performance of the algorithms is essential to avoid hallucinations that can prevent potential customers from using models in high-stakes environments where real dollars are spent.
Governance, transparency and explainability, enforced by true regulation, are key to giving businesses confidence that they can understand what AI is doing when missteps inevitably occur so they can limit the damage and work on improving the AI. There’s much to applaud in industry leaders’ first steps to create well-thought-out guards with real teeth, and I call for the rapid adoption of smart regulation.
Additionally, I would require any media (text, audio, image, video) generated by AI to be clearly labeled as “Made with AI” when used in a commercial or political context. Much like nutrition labels or movie ratings, consumers deserve to know what they’re getting into – and I think many will be pleasantly surprised by the quality of AI-generated products.
apps that kill
Hundreds of companies have sprung up in a matter of months, providing Generative AI, from the creation of marketing materials to the creation of new music and the creation of new medicines. ChatGPT’s simple prompt could potentially outperform the Big Data era search engine, but many other applications could be equally powerful and profitable across different verticals and applications. We are already seeing massive improvements in coding efficiency using ChatGPT. What else will follow? Experimenting to find AI applications that deliver a sea change in user experience and business performance will be key to creating feasible AI.
Companies that build their fortunes on this new class of technology will overcome these barriers to innovation. They will solve the challenge of continuously and cost-effectively building trust in AI while developing killer applications paired with sound monetization based on powerful underlying models.
Big data has gone through the same cycle of noise and nonsense. Likewise, it will likely take a few generations and a few missteps, but by focusing on the principles of Workable AI, this new discipline will evolve rapidly to create a platform for radical change that is just as transformative as experts expect.
Florian Douetteau is CEO of Dataiku.
Welcome to the VentureBeat community!
DataDecisionMakers is where experts, including data technicians, can share data insights and innovations.
If you want to learn more about cutting-edge insights and up-to-date information, best practices, and the future of data and data technology, join us at DataDecisionMakers.
You might even consider contributing an article your own!