AI is not yet as successful as it should be, so at the 2026 conference, 20 companies will reveal their move to an important new paradigm: hybrid AI.
Eric Siegel (with CHATGPT)
Originally published on Forbes
Having served as chair and keynote speaker at countless machine learning conferences over the decades, I have witnessed time and time again that event programs often signal emerging industry trends. This year, I’m chairing an event where dozens of companies are highlighting their move toward an important new paradigm: hybrid AI. Here, we discuss why the AI industry needs to go hybrid and the different ways companies are already doing so.
AI needs a breakthrough
So far, AI has not been as successful as it should have been. As the business world has come to understand, while many find AI useful, it falls far short of the enormous value promised. There are few genAI pilots reach production, Few machine learning models Expand.
The expectation that we will achieve astronomical values persists, but many are betting on unrealistic narratives about how we will get there. The core technology will improve and then the sheer suitability of the machines will overcome today’s lackluster track record. This has become the common story. AI experts and CEOs regularly proclaim that once LLM becomes “more intelligent” or ML models become “more accurate,” the technology will become seamlessly operational and deliver rich value to businesses and their customers.
Not so fast. There is no miracle solution at hand that is guaranteed to solve genAI’s stubborn reliability issues. Predictive AI also doesn’t have a reputation for being difficult to use.
Rather than betting on speculative and imaginary advances, I believe the solution will come in the form of a revolutionary paradigm shift: hybrid AI. Combining genAI and predictive AI, using each to address the other’s biggest weaknesses, multiplies the value provided by the technologies and core capabilities that already exist today.
How hybrid AI works
Predictive AI Addressing critical reliability issues in genAI. Rather than planning AI deployments based on hopes and prayers that genAI will become reliable enough to perform high-level functions autonomously, we need to recognize that humans will need to stay in the loop. With this in mind, the architecture of any ambitious genAI-driven system will be “predictive” in nature. Use predictive AI to target human intervention, but only in relatively high-risk situations where human intervention is most likely to be required. In this way, genAI can operate autonomously even in many cases. This approach delivers on a healthy part of genAI’s often bold promise of autonomy.
Similarly, genAI Addressing the complexity issue of predictive AI. Data scientists can develop robust models that put odds on who will click, buy, lie, or die, but to get the green light to deploy a model, business stakeholders must first develop a semi-technical understanding of what it means to operate such a model, that is, to systematically act on the case-by-case odds that the model calculates. LLM provides an important spoonful of sugar. Chatbots act as friendly experts, increasing understanding and thereby bridging the devastating technology and business divide that stalls most predictive AI projects. Additionally, LLM facilitates model development and provides visibility through vibecoding and deriving predictive features from unstructured data. Explaining ML model decisions To the general public.
Over the past year, I’ve provided several articles and presentations promoting this type of hybrid AI approach. Article on 5 ways to hybridize predictive and generative AI And one more thing Articles about two additional methods. I have argued that hybrid AI represents: Last hope before the AI bubble explodes.
But when I decided to dedicate the 2026 edition of Machine Learning Week, the conference series I founded in 2009, to hybrid AI, I wasn’t convinced that the obvious inevitability of a hybrid approach had yet taken root in the industry and had begun to materialize. Can I write an agenda for a two-day meeting?
Dozens of companies are using hybrid AI
A hybrid is happening. This year’s MLW conference will be held Hybrid AI 2026 – May 5-6 in San Francisco – and Alphabet We had no problem filling the program with presentations from numerous companies including; Schneider Electric, Spotify, State Farm, Twilio, Wynn Las Vegas.
We’re hearing a clear call to combine flavors of AI. “Today’s proliferating investment in LLMs is clearly inappropriate,” said Kirk Mettler, chief data scientist at IBM. Keynote on the fusion of LLM and enterprise machine learning. “Yet, it would be equally inappropriate not to leverage LLM to power your predictive AI projects. We found a sweet spot by marrying LLM with a long tradition of enterprise machine learning.”
He’s not the only one. As described in detail, meeting agendaenterprises are demonstrating hybrid AI in a rich set of use cases, including:
other companies A hybrid approach to ensure a much-needed “layer of trust” for genAI across use cases and application areas. The business world wants high-level functions to be fully automated, exemplified by the ambitious buzzword “agent AI.” Generally overpromising – Many such systems only have the potential to be ready for deployment with predictive guardrails. Salesforce presents on “”.When an AI agent goes out of control” is posted on the website. Predictively monitor humans. And we’ll kick off the event with a keynote on how genAI’s trust layer protection is expressed. Predictive AI’s new killer app.
Beyond the core conference topics, we also found that educational programs on hybrid AI were in demand. this Hybrid AI event Pre- and post-conference training workshops led by industry leaders james taylor and Dean Abbott.
Hybrid AI comes to the rescue
The time has come for businesses to start realizing the value of AI, which everyone talks about but few realize. The world needs it. Not only to improve AI and move us closer to realizing the promise for which so many have bet so much, but also because it’s all about value…efficiency improves the world’s prosperity. This is what technology is for.
We are excited to see this important industry direction come to fruition. Not only did we receive an overwhelmingly competitive response to our initial call for speakers on the subject of Hybrid AI to build our event agenda, but we are still continuing to receive a steady stream of speaker submissions, despite the deadline listed at the top of the submission form as 3 months in advance. Additionally, interest in events and feedback from seasoned experts continually reaffirms that the hybrid AI debate is resonant in the AI industry.
The world cannot afford to wait for speculative and potentially unrealizable advances that will achieve unprecedented levels of technological “smartness” and human-level capabilities. Instead, combining the two main characteristics of AI, predictive and generative, directly addresses the limitations of each and enables businesses to realize value today.
Access an overview of HYBRID AI 2026 and descriptions of each enterprise presentation. here (To attend the event in San Francisco on May 5-6, register (Prices reduced by March 20th). Disclosure: As the founding program chair, I am a partial owner of the Machine Learning Week conference series. – Includes HYBRID AI 2026 – And I will receive an honorarium as chairman.
About the author
Eric Siegel is a leading consultant helping companies implement machine learning and a former professor at Columbia University. He is the founder of the long-running Machine Learning Week conference series, the instructor of the highly acclaimed online course Machine Learning Leadership and Practice – End-to-End Mastery, editor-in-chief of The Machine Learning Times, and frequent keynote speaker. He wrote the best-selling book Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, and Die, which is used in hundreds of college courses. He also wrote The AI Handbook: Mastering the Rare Art of Machine Learning Deployment. Eric’s interdisciplinary efforts bridge the stubborn technology-business gap. At Columbia University, he received an Outstanding Faculty Award for teaching a graduate computer science course in ML and AI. He later served as a business school professor at UVA Darden. Eric has also published analytical and social justice editorials. You can follow him on LinkedIn
