Study Sees Dental Robots Sensing, Adjusting Mid-Surgery

Dental robotics is moving past implant guidance. A new review in the International Dental Journal looks at how AI-powered robots could sense, plan, and assist during dental procedures. This article covers what dental robots can do today, what is still experimental, and what the clinical evidence shows.

TL;DR

• Researchers propose combining AI, sensors, and robotics to assist during dental procedures.

• Clinicians already use robotic guidance for selected implant surgeries.

• Studies report high implant placement accuracy, with limited long-term outcome data.

• Fully autonomous dental treatment still needs safety and clinical validation.

What Is Dental Robotics, and How Does AI Fit In?

Dental robotics means a machine that physically moves or constrains an instrument in the mouth. In practice, that still means implant surgery almost exclusively. The robot holds the drill to a trajectory planned on CBCT, and the surgeon does everything else.

AI is a separate layer. Most dental AI never touches a patient. It segments CBCT volumes, traces the inferior alveolar canal, or flags caries on bitewings. Tools like these are covered in the guide to AI in dentistry. Several implant robots use AI during planning, such as automatic nerve tracing. They then execute that plan with conventional robotics. Nothing in the arm itself is learning.

That split is worth remembering when reading product brochures. "AI-powered robot" usually means AI-assisted planning plus a robot that follows instructions.

Here is how the main technologies compare:

What Makes Embodied AI Different From Conventional Dental AI?

Today's implant robots work from a fixed plan. The plan is locked before the first drill touches bone. The robot defends that plan, and the surgeon interprets everything else.

Any experienced implant surgeon reads the drill constantly. Resistance drops as the drill leaves dense cortical bone and enters cancellous bone. It changes again near the sinus floor or a lingual cortical plate. Surgeons adjust speed, pressure, and depth by feel, often without thinking about it.

Embodied AI would turn that tactile judgment into data. Force and torque sensors could register a sudden loss of resistance and stop advancing before a perforation. The system could then alert the surgeon or adjust the next step. That closed loop between sensing and movement is what the October review describes.

What Did the New October 2026 Research Find?

The review is titled The Prospect of Embodied Intelligence in Dentistry. Siwei Wang, Huancai Lin, and Liangyue Pang published it online in the International Dental Journal in October 2026. It is a narrative review. The authors drew on dental, medical, and engineering research available through July 2026.

The review proposes a framework built on four capabilities: perception, knowledge integration, planning, and execution. Together they form a loop. The system senses what is happening, checks it against patient data, decides the next move, and acts. Sensor data from that action feeds into the next cycle.

The authors map the framework onto implant surgery, endodontic microsurgery, orthodontic monitoring, and simulated treatment planning. They acknowledge that most supporting evidence is experimental, preclinical, or drawn from small clinical series.

The format matters when weighing the conclusions. A narrative review is a structured expert synthesis. It does not pool data or grade evidence the way a systematic review does. Its value lies in the framework, which gives researchers and manufacturers a shared vocabulary.

The framework also shows where current robots fall short. Today's implant systems handle perception of drill position and physical execution well. They stop at knowledge integration and real-time planning. A robot that tracks where the drill is, without sensing what the drill is cutting, covers only half of the loop.

How Are Dental Robots Used in Clinical Practice Today?

Implant surgery is the only area where dental robots see routine clinical use. Everything else sits between the lab bench and small first-in-human reports.

Robot-Assisted Dental Implant Surgery

The workflow will look familiar to anyone who has done guided surgery. It starts with a CBCT scan and a prosthetically driven plan. The difference comes at registration. The patient wears a splint or tracking array that links the CBCT coordinates to the actual jaw.

Registration decides much of the result. A tooth-borne splint on stable teeth registers well. Edentulous arches need bone-anchored arrays, which add a surgical step. Scatter from crowns and metal restorations can degrade the scan that registration depends on. A splint that rocks even slightly carries that error into every osteotomy.

During drilling, the surgeon holds the handpiece. The arm resists movement outside the planned position, depth, and angle. The surgeon still feels the bone, manages irrigation, and decides when to stop. Neocis Yomi works this way and was the first robot cleared by the FDA for dental implant surgery. The newer Yomi S received 510(k) clearance in October 2025.

Robotic Endodontic Microsurgery

Finding a root apex through intact buccal bone is one of the harder steps in apical surgery. Thick mandibular cortex, a nearby mental foramen, or the palatal root of a maxillary molar leave little room for error. A misplaced osteotomy means more bone removal and a slower recovery. Static targeting guides and dynamic navigation already help here. Robotic guidance has been tested in laboratory models and a few clinical reports. Robotic root canal treatment is not a routine service.

Orthodontic Monitoring and Smart Appliances

Orthodontists apply forces they cannot measure once the patient leaves the chair. Sensor-equipped brackets and aligners could report force levels and tooth movement between visits. That could mean earlier refinements and fewer surprises at the next appointment. These devices only monitor. No robot moves teeth or adjusts appliances on its own.

Experimental Tooth Preparation and Other Procedures

Restorative work is far harder to automate than implant drilling. An implant robot cuts rigid bone against a fixed reference. A crown preparation involves soft tissue, saliva, bleeding, a moving tongue, and subgingival margins. Each of those changes second to second. Research groups have built prototypes for automated tooth preparation and tooth autotransplantation. In 2024, Boston-based Perceptive reported a fully automated restorative procedure on a patient in Colombia. The system is not FDA-cleared, and no peer-reviewed clinical data accompanied the announcement.

How Accurate Is Robot-Assisted Dental Implant Surgery?

Accuracy is where robotic implant surgery has the most data. A 2026 systematic review and meta-analysis in the Journal of Dentistry pooled 27 clinical studies covering 983 patients and 1,546 implants. Robotic guidance was used for 1,018 of those implants.

The review measured how far each implant ended up from its planned position:

These numbers need clinical context. Many surgeons plan with a 2 mm safety margin above the inferior alveolar canal and about 1.5 mm from adjacent roots. A mean deviation under 1 mm sits comfortably inside those margins. Averages hide outliers, though. One implant 2 mm off at the apex can still reach a nerve or a sinus.

Angular deviation deserves the closest look. A small angle error becomes a larger apical error on longer implants. It also affects the restoration. A screw-access channel that exits through the facial surface of an anterior crown can force a cemented design or an angled abutment.

The results line up with earlier work. A 2025 meta-analysis of 16 clinical studies reported 0.69 mm platform, 0.72 mm apex, and 1.62° angular deviation. Accuracy also differed by tracking technology, which suggests the specific hardware matters as much as the concept.

Measurement method matters too. Most studies superimpose a postoperative CBCT on the original plan. That matching step carries its own error, and scatter around titanium blurs implant edges. Comparative data still favor robotic guidance. A 2025 meta-analysis in the Journal of Prosthetic Dentistry found computer-assisted workflows more accurate than freehand placement, with robotic systems the most accurate clinically. Study designs and robot types vary widely, so pooled figures are best read as a range.

Placement accuracy is a technical outcome. Implant survival depends far more on bone quality, primary stability, soft tissue, occlusion, and patient factors such as smoking and diabetes. No robot controls those.

Dental Robots vs. Traditional Implant Surgery: What Are the Differences?

There are four main ways to place an implant today. Freehand surgery relies on the surgeon's skill and judgment. Static guided surgery uses a 3D-printed guide made from the digital plan. Dynamic navigation tracks the drill on a screen in real time. Robotic guidance adds a physical arm that keeps the drill on the plan. The overview of guided implantology covers static and dynamic guidance in more depth.

Each approach has its own best fit. Freehand placement remains reasonable for wide, healed sites with good visibility in experienced hands. Static guides work well for cases planned in advance. The sleeve and longer drills need generous mouth opening, though, and second molar sites are where guides most often fail. Sleeves also limit irrigation and make it hard to change course once drilling starts.

Dynamic navigation and robotic guidance drop the sleeve and allow the plan to change mid-surgery. Navigation asks the surgeon to watch a screen instead of the field, which takes real practice. Robotic guidance moves the constraint back into the handpiece, so the surgeon's eyes can stay in the mouth.

Immediate placement is a good test. In a sloping extraction socket, the drill tends to skid toward the thinner wall. A rigid arm can resist that deflection more reliably than a hand. Full-arch cases are another. Bone reduction and several angled implants benefit from a system that holds every position to one plan.

No approach wins in every case. Results depend on the system, the operator, the site, and how each study measured accuracy.

Which Dental Robots Are Available in 2026?

Only a handful of dental robots have regulatory authorization anywhere. Most activity is in implant surgery, with the US and China as the main markets.

Neocis Yomi is the main system in the US. It is a robotic navigational system for implant planning and intraoperative guidance. The dentist plans in YomiPlan software from a CBCT scan, and the arm provides haptic guidance during the osteotomy. According to Neocis's indications for use, Yomi is intended for partially and fully edentulous adults who qualify for implants. Its cleared uses also include guided bone reduction. The company launched the next-generation Yomi S in late 2025 and reports more than 100,000 osteotomies on the platform.

China has moved fastest on more autonomous designs. Remebot received authorization from China's National Medical Products Administration as an autonomous implant robot. Several other Chinese systems appear in the clinical literature, though few have published regulatory details in English.

Here is how the main named systems compare:

None of these companies publish pricing. Before taking a demo at face value, a practice should ask a few direct questions. What imaging and splints does the system require? How long does setup take on a typical single implant? Which implant brands and drill kits does it support? What published clinical data exist for this exact model? A research prototype should never be read as a product a practice can buy.

Can Dental Robots Perform Procedures Without a Dentist?

No dental robot is cleared to treat patients without a dentist. Autonomy comes in levels, and the difference between them matters:

• Passive assistance means the robot holds or positions something while the dentist works.

• Active guidance means the robot physically limits where the dentist's instrument can go.

• Supervised automation means the robot performs a step on its own while the dentist watches and can stop it.

• Autonomous operation means the robot plans and treats without a clinician directing each step.

Current US systems sit at the active guidance level. The dentist plans the case, approves the trajectory, monitors the procedure, and remains legally responsible. That also means the dentist owns the plan. If the plan places an implant too close to the canal, the robot will follow it faithfully.

Lab demonstrations happen on fixed models with no bleeding, saliva, or tongue. Patients swallow, move, and react to discomfort. Any robot working in the mouth needs reliable tracking of patient movement, a fast emergency stop, and a clear handover to the dentist. Those safeguards have to be proven in patients.

What Are the Benefits and Risks of Robotic Dentistry?

The main benefit is consistency. A well-registered robot places the implant where it was planned, whether the surgeon has placed fifty implants or five thousand. That narrows the difference between operators, especially in posterior sites with limited access. It also makes flapless surgery more predictable in suitable cases. Flapless is not right for every site. It removes the surgeon's view of the buccal plate and the chance to manage keratinized tissue.

The broadest look at the clinical record so far is a 2026 systematic review in the Journal of Prosthetic Dentistry. It covered 36 clinical studies published between 2020 and 2026. The authors found robot-assisted implant surgery safe and highly accurate, with favorable short-term outcomes. They also noted longer setup times and limited long-term follow-up.

The risks are mostly technical and practical:

• A sound execution of a poor plan is still a poor result, and the robot cannot judge the plan.

• Registration errors shift every osteotomy if the splint or tracking array moves.

• Tracking failures can interrupt guidance mid-procedure.

• Patient movement needs constant compensation, especially in sedated or anxious patients.

• Equipment malfunction requires a fallback plan and solid freehand skills.

• Extra setup steps slow a practice down during the learning curve.

Claims of fewer complications or faster surgery need comparative trials to back them up. For now, the evidence supports accuracy and short-term safety.

How Much Does Robotic Dental Surgery Cost?

Cost questions come from two directions, practices weighing a purchase and patients comparing quotes. Neither has a simple public answer.

For practices, manufacturers do not publish list prices for dental robots. Quotes are negotiated and often bundle hardware, software, service, and training. The purchase is only part of the cost. Practices also pay annual service and software fees, per-case splints or arrays, extra CBCT scans, and chair time during the learning curve. Implant volume decides the math. A practice placing a few implants a month will struggle to justify a robot. A high-volume surgical or full-arch practice faces a different calculation. Some owners also count referral and marketing value, which is real but hard to measure.

For patients, the price of a robot-assisted implant varies widely between clinics. Some practices charge a premium for the technology, and others absorb it. Insurance benefits generally apply to the implant procedure itself, and the use of a robot rarely changes what a plan pays. Patients should ask for an itemized estimate and check their benefits before treatment.

What Prevents Wider Adoption of AI-Powered Dental Robots?

The October review lists several obstacles between today's robots and true embodied AI. Most are about data and systems that do not talk to each other.

Data is the first problem. A robot that adapts its movement needs to learn from thousands of recorded procedures, with force traces, drill positions, and outcomes. That kind of dataset barely exists in dentistry. Most practices record no intraoperative data at all, and what exists sits in proprietary formats. Hardware and software from different vendors rarely connect. A surgeon who plans in one program may need another to run the robot. Clinicians also need to see why a system made a decision before trusting it near a nerve. Regulators have few precedents for a device that changes its own behavior during treatment.

A separate review published October 9, 2026 looked specifically at AI in implant surgery. It identified data inconsistency, closed hardware systems, and limited prospective validation as the main obstacles.

Workflow is the barrier most practices feel first. A robot changes imaging protocols, splint fabrication, scheduling, and staff roles. Early cases usually run longer than the surgeon's normal workflow. Broader trends in dental AI adoption show where practices stand with AI tools before adding robotic hardware.

The next step is prospective, multicenter trials with longer follow-up. Those studies will decide whether these systems improve outcomes patients can feel.

How Could Embodied AI Change Dental Treatment Planning?

A robot executes a plan with great precision, including a bad plan. Planning quality therefore matters even more with robotics. That means prosthetically driven implant positions, careful tracing of the canal and sinus, and intraoral scans merged with the CBCT for accurate tooth and soft tissue surfaces.

Digital twins are one proposed way to bring these together. A digital twin is a virtual model of the patient's mouth. Dentists could rehearse a procedure on it, test different implant positions, and see how a plan behaves before touching the patient. The robot would then work from a plan that has already been checked.

Good planning starts with good records. A structured dental treatment plan keeps findings, options, and decisions documented in one place for clinician review. Guidance on preparing a dental treatment plan covers the clinical side. Neither tool controls robotic equipment. Their role is clear documentation, which any future robotic workflow will depend on.

What Comes Next for Dental Robotics?

Progress over the next few years will likely be gradual. Most research is aimed at making today's implant robots more responsive before moving into new procedures.

The main research directions include:

• More responsive force and movement sensing

• Tighter integration with CBCT and intraoral scans

• Better surgical registration and patient tracking

• Clinical trials of robotic endodontic microsurgery

• AI decisions that clinicians can inspect and understand

• Longer follow-up of implant survival and patient outcomes

No regulator has set a timeline for autonomous dental treatment in the US, and company forecasts are estimates. The signals worth watching are FDA clearances, peer-reviewed clinical trials, and long-term outcome data.

Bottom Line

Dental robotics is already part of implant surgery. FDA-cleared systems like Yomi hold the drill to a CBCT plan while the surgeon stays in control. Placement typically lands within a millimeter of that plan. The weak points are registration, plan quality, and long-term outcome data. The October 2026 review describes robots that sense bone and adjust in real time. That remains research. For now, a robot is only as good as the plan and the surgeon behind it.