AO or AI? Reflections on Progress and the Power of Human Association

If given the option of AO or AI the simple answer is to choose ‘C’, both AI and AO, and avoid the dichotomous nature of the question. By not allowing the ‘both’ option we can explore the deeper meaning and role of each entity for us as spine surgeons.

For historical context, the origin story of AO cannot be retold enough: In 1958 a group of surgeons changed surgical education by congregating in person in a hotel near the Swiss alps and proceeded to formalize modern musculoskeletal trauma care in an interdisciplinary multinational fashion with their pragmatic reproducible ‘evidence -based’ (yes, they applied that term way before it became fashionable ‘Medspeak’ in the 1990’s). From an initial nucleus of General, Trauma and Orthopaedic surgeons these academic and private practice Swiss, Austrian, German, Belgian and French surgeons rapidly gained traction with the unprecedented idea that surgical education should not just be a domain of local medical schools or national or specialty societies but would be best fostered by a group inspired by the common idea of creating a greater good – something that all patients and practitioners could benefit from. ¹

‘Arbeitsgemeinschaft Osteosynthesefragen’, in short ‘AO’, the literal question of ‘how to get bone to heal’, became a somewhat unlikely nucleus for an expansive evolution of future education of surgeons, research and technique development transcending borders, continents and surgical specialties.

The true reason for the success of AO was less the subject matter of how to ‘improve fracture care’ but rather the founding members’ unprecedented ability to transcend prior man-made boundaries by applying a universally appealing pragmatic approach of analysis, description, then classification followed by application of care pathways using the most sensible scientifically proven techniques of their times and then insisted on following this up with formalized outcomes tracking. This simple but hitherto untested approach would then in turn foster an ongoing standardized pragmatic educational effort delivered in a standardized format that heavily emphasized interactive learning rather than pure lecture delivery. The magic of ‘fireside chats’ – improvised, unstructured and real, at the end of a long day of meetings will not be forgotten by any of those in attendance. As simple as this progression of knowledge creation and dissemination may sound today the consistent application of principles applied to common problems across previously established human created restrictions delivered by motivated participants has been a ground-breaking success story for more than 75 years. Their four founding principles have stood the test of time: 1. Documentation of all patients; 2. Development of new implants and instruments; 3. Research of fracture healing and tissue cultures; 4. Teaching of osteosynthesis techniques. It is pretty remarkable that these common sense principles would serve as foundation for a multispecialty and multidimensional expansion with Basic Scientists, Epidemiologists, Veterinary surgeons, Oromaxillofacial, Plastic Surgeons, Hand Surgeons and Neurological Surgeons, Pain Specialists to more recently Orthopaedic Adult Reconstructive and Sports Medicine surgeons joining the ranks of this movement and evolving it into a global behemoth of about 8,000 faculty teaching at 900 courses per year for 65,000 attendees. ² On the AO Spine International side an important fifth dimension was added to this movement 15 years ago in form of a knowledge creating entity simply titled ‘Knowledge Forum’. This entity presented an unprecedented opportunity for colleagues across all regions and specialties of the world to become engaged in problem focused subspecialty research independent of corporate sponsored ‘study groups’ and thus help advance Spine care along five subspecialty interest groups. All this represents an impressive and sustained investment of time, methodologic, logistics and financial support organized by the AO Foundation and its subdivision of AOSpine International as well as its members around the globe. And just as this system has gained impressive momentum in terms of research output the very real advent of vast information technology resources at the fingertips of every connected human being (so their government allows such) calls into question how we learn, analyze and teach.

The advent of AI has literally transformed our potential of knowledge review and analytics overnight. To think that Chat GPT was created in 2022 as a loosely organized noncommercial collaborative research and study effort by the name of Open AI, and Claude, created by another hitherto unknown entity called Anthropic, started in 2023 is a mindboggling reflection of just how seismic the advent of publicly useable Artificial Intelligence has been and has become a core part of our lives. ³ Traditional major IT companies and other startups are all frantically competing for preeminence in this transformative space with profound ethical and frank power-based implications. Both Open AI and Anthropic are going to become publicly traded in the near future and probably set new records on market valuation. National and international regulatory efforts as well as attempts to meet the vast energy resources needed to fuel these enterprises are lagging well behind the continuously evolving tech revolution taking place in front of our very eyes. In our field of Medicine the arrival of a dedicated program called ‘Open Evidence,’ (OE) founded in 2022 by two entrepreneurs, Dan Nagler and Zach Ziegler, is of particular interest as this is a large language search engine specifically created to help medical professionals better analyze and stratify the ever-growing mountain of medical knowledge. ⁴ This company gained notoriety for having allegedly scored 100% on the US Medical Licensing Exam in 2025.after scoring ‘only’ 90% on this notorious exam two years earlier.^5,6

Literally overnight and seemingly propagated by word-of-mouth this search engine has become a fixture on the handhelds and desktops across North America due to the quality and depth of its responses. OE reports 760,000 US physicians and 18 million consultations/month. ³ While similar results could probably be obtained by training other large language models (LLM’s) this particular system offers more curated and properly referenced responses at moment’s notice. Interestingly access to Open Evidence remains blocked in countries such as Western Europe as of the Summer of 2026. (personal observations).

Inevitably such vast knowledge resources, if released to the public by their respective authorities, are bound to transform how we learn and practice medicine. Current LLMs are limited by applying strict phrase recognition and extraction methodology for their abstractions. Predictably they use key words, follow an evidence pyramid scheme as originally proposed by Guyatt et all and also weigh study size and ultimately pull out anything published online into their summaries. ⁷ This leads to an output that usually starts with Systematic Reviews and where applicable Metanalyses, and then hierarchically emphasizes prospective randomized trials regardless how sensible or well performed they were. Adjudication of factors such various forms of bias, regional culture of care and sensibility of data beyond a superficial layer of reasoning are readily elusive to these search engines and in absence of other data mere opinion pieces of commercial entities may rise to the status of decision influencing play makers even in absence of comparison studies. Well publicized examples of fantasizing references and presenting unacceptable forms of chauvinism have lead to calls for greater regulations of AI and in turn lead to ongoing efforts at refinement of search engine extraction and valuation methods.^8,9 Despite such flaws the bandwidth and sophistication of search engines is exponentially growing in leaps and bounds. The only question will be who asks the best next-gen questions, can train their platform to suit their needs best and can find meaningful ways to apply the output to meet their goals.

Lectures of the near future will probably no longer consist of a PowerPoint presentation painfully compiled by an invited presenter, but scan patient images with key variables into an existing search engine along with the request to present treatment options and their pertinent pros and cons based on key references. The audience interactions would consist of modulating patient variables and discuss the various results predicted by an AI platform. A particular high point of such a future presentation could be to feature a battle of various AI vendors against one another (“Tonight’s Battle of the Titans: Gronk versus Claude assess Degenerative spondylisthesis management!!”) and have the audience score them for making the best points.

As to knowledge creation merging the enormous patient data pools of electric medical records behemoths such as Epic through filters and refined ‘slice and dice’ tools could lead to immediate outcomes reporting for common as well more exotic conditions and surgeries. ¹⁰ The only present shortcomings to offering real life data reporting, for instance of surgical complications, are currently perpetuated by inconsistent coding quality, lack of structured outpatient outcomes data integration and self-imposed proprietary hospital system restrictions. The data is already available now and all is needed are organizational directives to address these known limitations and there would be vast real time data reports available on any imaginable topics. Just imagine the question of the incidence of footdrop after isthmic spondylolisthesis surgery could be answered at moment’s notice with actual data reporting relative to surgical techniques, clinical outcomes, other complications, patient factors and yes – surgeon factors! The surgical registries of the past perennially suffered from underfunding over time, selection biases paired with complex gathering methodologies and perennial incomplete data sets. With advanced Electronic Medical Records composite data base sets are available and will be used as real-life decision-making tools instead of relying on published studies generated in far-away locations by distant surgeons operating under their individual practice circumstances, that may not at all apply to local standards.

So, if past medical knowledge in neatly formatted presentation is immediately available through increasingly powerful LLM based search bots and the current power of large EMR conglomerates gets unleashed by progressive integration and structured analytic tools, where does this leave traditional large physician collaboratives, such as illustrated in the AO and the Knowledge Forum example above? Will traditional lecture eminence-based lectures and ultimately anecdotally based teaching formats become obsolete in the face of the advent of this AI fostered tech revolution? Will the enormous individual and organizational investments of time and resources for in person meetings and large research cooperatives like the Knowledge Fora become simply unnecessary?

In short – no! And the reason lies in a single word, a word that describes the superiority of an in-person gathering as information exchange and knowledge generation event over any AI resource. The word in question is: Alive. Meetings – if well done, are ‘alive’. Aside from the formal (PowerPoint) based lectures and presentations they provide a meeting place for a cross-section of humans of all imaginable age groups, sexes, belief systems and experiential back grounds in an organized yet random jumble of interactions and exchanges. And it is through this unpredictability that new ideas can arise, prior concepts get modified and new collaborations emerge. The more a meeting can plan for such unplanned interconnections to happen the better it is. While this in theory can apply to any professional gathering– and there are definitely too many of them – AO meetings do stand out. The recently concluded Global Spine Congress organized by AOSpine International in Istanbul end of May 2026 is a great case in point. With over 2,800 physician attendees present despite a major ongoing conflict in the region, it was joy to see the energies set free by cross-cultural and trans-age interactions. The key common bond was a universal fondness for AO, a de facto fellowship of surgeons from every conceivable background across the globe. This common cause and background of AO lead to an effortless lingua franca to be spoken, from the impressive ‘Youth Club’ of a very well-presented younger Spine surgeon generation to a ‘Women in Spine’ section to presentations from venerable Spine and Subspecialty societies all over the world. And the Industry sections, which are increasingly ignored in larger meetings, were buzzing like a bazaar with open walled meeting places providing vocally presented exchanges of ideas. Throughout the day meeting spaces were filled with human interactions from morning coffees to host dinners in the evening and the resultant interactions were impressively vibrant. This congress was ‘alive’ and a clear remonstration of the power of human interaction. It is hard to fathom any digital environment recreating such a synaptic connectivity. The fellowship of mankind was on full display and left all participants energized and inspired.

So, the answer to the question posed at the onset of this article is clear – the fellowship of mankind represented by an overarching idea like ‘AO’ will supersede the emerging AI wave effortlessly as it fulfills its participants with a sense of purposeful bonding. And AI remains what is was designed to be – a method to support work of humans and neither cause nor content. That said the option ‘C’ – AO and AI together – remains the obvious choice. AO courses of the future will need to evolve to integrate AI in their teachings and delivery. The future is bright and AO and AI will be an amazing combination to look forward to.