A world with autonomous robots and embodied AI could be one filled with excitement. By 2030, those two types of technology will have evolved significantly and will make a significant difference in how we live and work.
An Embodied AI is an integration of artificial intelligence and a body or form that can interact with the physical world. An embodied AI incorporates sensory information from the outside world into its processing and action, based on what it has learned.
As time progresses, autonomous robots will be used for both simple tasks (such as cleaning) and complex tasks (such as surgical procedures). Autonomous robots have the potential to dramatically change how many different sectors of society function (for example, health care and manufacturing).
Advances in AI technology will improve robots’ decision-making capabilities. Improving robots’ decision-making abilities will enable them to think more adaptively. Additionally, advances in AI will enable robots to perform their tasks more efficiently.
The next decade (by 2030), will be exciting in terms of technological innovations but also present difficult ethical and regulatory issues. How societies choose to address these issues may help define a new era of technology.
Summary
By 2030, the development of autonomous robots and embodied AI technologies will shift from niche applications to ubiquitous components of everyday infrastructure across many sectors and parts of our lives. The integration of AI with a physical body (embodied AI) is projected to greatly enhance the ability of machines to perceive their surroundings, make decisions, take action safely and in real-time, and ultimately allow for much more natural ways to interact with humans and the ability to adapt quickly to changes in their environment.
Advances in sensor technology, actuators, machine learning algorithms, and natural language processing will enable future generations of robots to be far more accurate, capable, and collaborative, which will support the types of tasks found in everything from warehousing/picking/quality inspection in manufacturing to logistics within hospitals and care-giving services in homes.
Impact will likely be greatest in healthcare, logistics, manufacturing, transportation, and education, but it can be seen across other areas of society as well. The article points out that what will drive this widespread impact is systems that create efficiency by reducing or eliminating repetitive work, improving reliability, and enhancing the capabilities of existing human teams by using autonomous robots as augmentation tools rather than as replacement tools.
However, there are significant barriers to achieving large-scale use of these technologies, including ensuring safety and accountability in autonomous decision-making; protecting individuals’ right to privacy regarding the data collected; and mitigating negative impacts on employment opportunities and public trust.
Thus, the authors frame the path forward to 2030 as both a technological opportunity and a “governance test” requiring designers to design responsibly; regulators to develop regulatory frameworks that address concerns about the uses of physical AI at large scales; and users to ensure that physical AI is deployed in a way that is centered around people.
What Is Embodied AI? Understanding the Basics
Embodied AI combines the elements of an intelligent system (intelligence) and physical structure (body). Embodied AI includes integrating artificial intelligence into a physical body that can interact in the real world. The use of embodied AI enables machines to mimic human interaction and enhance their functionality in many ways.
The concept of embodied intelligence revolves around a few core components:
- Sensory Input: Gathering data from the environment through sensors.
- Processing: Analyzing data to make informed decisions.
- Physical Action: Executing actions based on decisions.
The integration of the two elements allows a machine to function more naturally in its environment. Unlike traditional AI, which lacks a physical form, embodied AI uses a machine’s physical form to enhance its ability to engage with humans and the environment.
Embodied AI vs Traditional AI vs Physical AI
| Feature | Traditional AI | Embodied AI | Physical AI |
|---|---|---|---|
| Physical Presence | No | Yes | Yes |
| Learns from Environment | Limited | High | High |
| Uses Sensors | Minimal | Extensive | Extensive |
| Real-World Interaction | No | Yes | Yes |
| Example | Chatbots | Service Robots | Autonomous Humanoid Robots |
| Decision Making | Data-based | Environment-based | Environment+Action-based |
Example
A customer service chatbot can answer questions, while an embodied AI robot can physically guide visitors through a building.
Source: Stanford Human-Centered AI (HAI)
Link: https://hai.stanford.edu
AI integrated into physical structures will be much more prevalent and flexible through 2030, with significant potential to impact many areas, including health care and transportation.
The Rise of Humanoid AI in Healthcare Logistics and Manufacturing
Embodied AI enables intelligent systems to learn, interact, and operate in physical environments
Embodied AI refers to smart software packages attached to a “body” (e.g., a robot or vehicle) that can perceive and react to its environment. They differ from traditional digital models in how they learn. Instead of simply processing data, Embodied AI learns through an iterative cycle of sensing, reasoning, and acting (i.e., vision, hearing, LiDAR, and touch sensors collect sensory information; algorithms process this information; motors and Actuators physically carry out actions).
The ability of Embodied AI to receive physical feedback enables the system to test hypotheses, correct mistakes during execution, and improve over time by operating in dynamic environments where objects move, lighting conditions change, and humans behave unpredictably.
They also require world models: a mental map of the space, object characteristics, and the probable consequences of taking some course of action. For example, if a gripper slips or a door is heavier than anticipated, the controller adjusts the force applied to the gripper and the path taken to reach the destination within milliseconds.
Many modern methods combine Reinforcement Learning (RL), Imitation Learning (IL) based on human demonstration examples, and Classical Planning (CP), after validating these behavioral patterns in simulated and pilot-tested environments. By doing so, Embodied AI developers can reduce uncertainty about how their product will perform and help meet Regulatory and Safety Standards across all environments, including those in which Robots and Humans coexist.
Since Embodied AI must be constrained by factors such as gravity, friction, energy limitations, and safety rules, it extends the capability of knowing what to do to the ability to do it. Examples of how Embodied AI operates include:
In Warehouse environments, Embodied AI can travel down aisles, recognize products, and manipulate objects with varying geometric dimensions. In Hospital Environments, Embodied AI can supply necessary materials and equipment to medical staff while avoiding pedestrians. In Home environments, Embodied AI can clean rooms, pick up requested items, and respond to verbal commands such as “Pick Up My Charger”. In Factory environments, Embodied AI enables Collaborative Robots to adapt to new components and production workflows without extensive retraining.
The ultimate potential for Embodied AI is to provide Scalable Autonomy: i.e., Machines that can learn to accomplish multiple tasks by combining Simulation-Based Models, Real-World Experience, and Human Feedback. As sensors continue to decline in cost and models become more computationally efficient, running on mobile devices, we expect that Embodied AI will become increasingly important in providing Intelligent Systems the capability to Learn, Interact, and Operate within Physical Spaces – exactly what Embodied AI was designed to do.
Example
An autonomous cleaning robot using embodied AI navigates a grocery store’s narrow aisle late at night during the night crew’s restocking. The robot uses both depth and tactile sensing to detect when a glass jar has been knocked onto the floor, slow down, and take an alternative route to prevent the spread of broken shards. Upon reaching the spill area, the robot transitions from tool mode to use a vacuum head to remove fragments and a mop module to clean up residue.
If a worker enters the robot’s path, it will pause and request permission before proceeding. Through continued operation within the store, it will develop knowledge specific to that store layout and also where spills commonly occur.
AI Technology: AI technology uses advanced algorithms and data to automate tasks and improve decision-making
The term “AI Technology” refers to a class of systems that utilize advanced algorithms and vast amounts of data to automate numerous processes and improve decision-making. The fundamental elements of this technology include combining machine learning, deep learning, and statistical modeling techniques to identify patterns and provide recommendations faster than traditional software systems.
Once an organization implements AI Technology, it generally integrates it with real-time data feeds (e.g., transactions, sensor readings, images, text messages) so that the AI Technology’s outputs are continually updated as conditions change.
One of the greatest strengths of AI Technology is its ability to perform automation at scale. For example, in customer service applications, AI Technology may be used to triage customer inquiries, generate response templates for simple queries, and automatically escalate complex issues to human customer service representatives. Similarly, in healthcare applications, AI Technology may assist clinicians in detecting abnormalities in medical images or identifying potential patient risks. Also, within financial institutions,
AI Technology has been applied to detect fraudulent activity, assess an individual’s creditworthiness, and optimize investment portfolios by continuously learning from historical behavior. Lastly, AI technology can be used in operational settings to improve forecasting models, optimize scheduling, and develop more accurate supply chain planning models by converting unstructured data into meaningful insights.
In addition to providing automated solutions, AI Technology also offers enhanced decision-making alternatives rather than replacing human decision-making. As such, many AI-based systems are developed around “human-in-the-loop” workflow concepts, in which the AI technology presents recommendations, but the person retains ultimate authority.
Consequently, when the outcome of an action could have implications related to public safety, equity/fairness, or regulatory compliance, these factors become extremely relevant during the design phase of responsible implementation, which includes considerations for monitoring, transparency/explainability/bias detection.
Example
A local hospital implements an Artificial Intelligence (AI) tool to help alleviate congestion in its Emergency Department. This AI tool can collect real-time information from a variety of sources, including ambulance arrival times, laboratory result turnaround times, available beds, and the percentage of clinicians currently working. Based on this real-time information, the AI tool creates predictions regarding patient flow over the next six hours.
The AI then provides recommendations on whether staff should be adjusted and which imaging slot will most likely be needed for potential stroke victims by identifying symptom-related entries made during the triage process. If there is a delay in obtaining laboratory test results, the AI tool will automatically send stable patients to an alternate testing location and indicate those patients ready for discharge for physicians to review prior to release.
Although the clinical team makes the ultimate decision, the AI tool continually updates its forecast. As a result, the hospital has successfully reduced waiting times, avoided crowding, and maintained operational efficiency during periods of increased demand.
Physical AI: Physical AI combines artificial intelligence with robotics to perform real-world actions autonomously
Physical AI is the integration of robotics and artificial intelligence, enabling machines to perform autonomous actions in the real world.
While software-based systems do not have to contend with the messiness of the real world as they make decisions about how to interact with their environment, Physical AI must be able to perceive its environment, determine safe motion plans to take action within that environment, and then execute those plans through use of motors, grippers, wheels and/or legs—while also being responsive to changes in the environment due to the presence of humans, objects and changing environmental conditions.
Therefore, Physical AI includes three fundamental components—sensing (vision, lidar, depth, touch), decision-making (planning and learning), and control (precision movement)—which function together continuously as an integrated system.
One of the most significant challenges faced by Physical AI is achieving reliability in a physically constrained environment. Uncertainty is introduced into the system by gravity, friction, sensor noise, limited power/battery life, and imperfections in the hardware itself. These uncertainties require Physical AI to learn from both simulated environments and real-world practice.
Often, these types of Physical AI systems combine reinforcement learning and imitation learning from human demonstration data with traditional robotic algorithms such as classical robot control. The benefits of this type of hybrid approach include the ability of Physical AI to generalize what has been learned during training to new rooms, new objects, and previously unseen edge cases without needing to be continually programmed.
Physical AI provides a common thread across many types of robotic applications, including robots that perform picking/packing functions in warehouse settings; robots that travel hospital corridors to deliver medical supplies; and robots that perform inspection functions, such as monitoring wind turbine blades and pipeline integrity.
Additionally, Physical AI enables adaptive automation in manufacturing, where robots can handle diverse product configurations, detect defects, and work safely and collaboratively alongside human employees. Finally, Physical AI is used in agricultural robotics, such as crop monitoring, weed detection, and applying treatment products to crops, achieving higher accuracy and lower waste than previous methods.
The term Embodied AI is very similar; however, it places a strong emphasis on creating intelligent systems whose abilities are grounded in bodies that allow them to sense and act on their environment. On the other hand, Physical AI focuses specifically on providing a robotics-based implementation stack that enables real-world autonomy.
Overall, while there is some overlap between the two areas of research (i.e., embodied cognition and autonomous robotics), they represent complementary directions for advancing the development of increasingly autonomous technologies capable of operating outside controlled laboratory settings. With advancements in sensors, more efficient execution of models on devices themselves, and increased capabilities in machine learning algorithms, Physical AI will continue to proliferate into home environments and urban spaces where it will be capable of performing more complex tasks safely, responsibly, and effectively.
Example
The physical AI Inspection Rovers are used by a solar farm to inspect the panels at night, using autonomous rovers on each row. Using both thermal imaging and computer vision, the rover detects hot spots on the panel, cracks in the glass, loose or damaged wiring, and other issues. Once the rover identifies an issue with one of the panels, the Physical AI System will determine a path for the rover to take to get close to the affected panel without damaging any cable, and once in position, use a small mechanical arm to attach a temporary identification tag.
Once completed, the Physical AI System will upload precise GPS Coordinates, images of the faulty area, and a severity rating so the Technicians know which areas require attention as soon as possible, at sunrise. Additionally, if the terrain is wet/muddy, the rover can adjust its traction by adjusting the torque of its wheels and alter its course to prevent becoming mired.
What Is Cloud Robotics and How It Work
Embodied Intelligence: Embodied intelligence allows machines to understand, adapt, and respond through physical interaction with their surroundings
The ability of machines to “know” things as they interact with their surroundings through sensors and movement is referred to as embodied intelligence. Instead of using only abstract information about what has been sensed, embodied intelligence develops because of how a system perceives the world – for example, by seeing the same object from various angles, sensing resistance when trying to push open a door, or making adjustments to keep one’s center of gravity while walking along uneven terrain.
The embodied nature of the system enables it to understand context and respond appropriately in uncertain situations, so embodied intelligence is especially important in robots expected to operate safely around humans, fragile objects, and variable environmental conditions.
As embodied intelligence occurs due to sensor input (e.g., vision, depth, touch, force, etc.), algorithmic processing of those inputs into a situational model of the environment and then outputted as action (and then later as learned outcomes based on responses to that action), and because the environment reacts back to the robot, embodied intelligence relies heavily on adaptability.
Therefore, if the robot loses its grip on something or if an unexpected obstacle arises, the system will immediately update its plan of action. That is why embodied intelligence is at the core of developing robots capable of operating safely around people, with fragile objects, and in a variety of unanticipated ways.
Many modern embodiments of embodied intelligence are created by blending together reinforcement learning, human demonstration-based imitation learning, and motion planning. While simulation provides a method to increase scalability during training, real-world experimentation combined with human feedback increases the robustness of the embodiment. Through continued use and learning over time, embodied intelligence may enable even more natural user interfaces (i.e., machines that can be instructed with phrases such as “place the cup on the top shelf” and navigate around crowds) and enable machines to act in concert with humans.
While there is no single agreed-upon definition of embodied artificial intelligence (AI) among researchers today, many refer to systems developed under this paradigm as embodied AI. When properly implemented, embodied AI uses embodied intelligence to transform perceptions into reliably executed actions, rather than simply predictions. For example, in warehouse settings, embodied intelligence enables effective picking and sorting of a variety of products. In healthcare settings, it facilitates delivery and assistance-type tasks. At home, it enables cleaning up messes, retrieving objects, and navigating safely.
Example
The robotic assistant, while assisting a resident with mobility issues who is using a walker in their living space, exemplifies “embodied intelligence” by adapting to the resident’s movements. The resident tends to pause at the doorway and use the walker for assistance, so the robot has learned to maintain a safe or respectful distance and position itself slightly off to the side to avoid blocking the doorway.
When instructed by the resident to bring her sweater from the armchair she sits in, it first scans the area, identifies the chair, then gently picks up fabric (without dragging) when touching down with its hand to lift and carry the sweater — adjusting its grip based on what it feels through touch feedback.
Key Technological Advancements Shaping 2030
The advancement of AI technologies toward 2030 has been tremendous. As a result of continued innovation in machine learning algorithms, there are now better, faster, and more accurate versions. Thus, as these new machine learning algorithms evolve, they increase the cognitive ability of autonomous systems.
As recently developed sensor and actuator technologies improve, so too does the overall functional ability of machines (robots). These technologies enable robots to understand their surroundings more accurately, resulting in improved task completion. As a result of these developments, robots will be able to perform multiple functions in numerous settings.
By 2030, it is expected that AI systems will have greatly increased in their ability to make decisions. As a result, AI will be capable of solving complex problems. Driving factors behind these changes include:
- Advanced Algorithms: Enabling better learning and adaptation.
- Enhanced Sensors: Providing detailed environmental insights.
- Refined Actuation: Improving precision and control.
AI and Robotics Market Growth Toward 2030
| Market Segment | Estimated Market Size by 2030 |
|---|---|
| Artificial Intelligence | $1.8+ Trillion |
| Robotics Market | $200+ Billion |
| Autonomous Mobile Robots | $18+ Billion |
| Industrial Robotics | $80+ Billion |
| Healthcare Robotics | $45+ Billion |
Key Statistic
The global AI market is projected to exceed $1.8 trillion by 2030, driven by automation, robotics, and intelligent systems.
Source: Grand View Research
Link: https://www.grandviewresearch.com
Advances in Natural Language Processing (NLP) will also enable humans and machines to communicate more smoothly. This will lead to easier human/machine interaction and will ultimately create an even simpler process for working together. Based on the current rate of development, we can expect many exciting developments in the years leading up to and beyond 2030, which will revolutionize how industry operates and how our lives function.
Real-World Applications: Where Embodied AI Will Make an Impact
By 2030, embodied AI is expected to revolutionize a variety of industries. In healthcare, robots powered by artificial intelligence will be assisting doctors and surgeons with complex surgical procedures, enhancing precision and outcomes. They will also help handle repetitive tasks and allow medical professionals to focus on critical patient care.
In logistics, autonomous systems are expected to streamline supply chain operations. Through artificial intelligence-driven robots, warehouse inventory can be managed and shipment process times reduced. This efficiency is expected to result in cost savings and lower carbon footprints for the sector.
Manufacturing will also see substantial impact from these advances; assembly lines and quality control will be handled by robots, which are expected to yield higher production rates and more accurate results. Beyond the industry, embodied AI technology will also improve personal lives by providing smart home assistants that perform household responsibilities.
- Healthcare: Surgical assistance and patient monitoring.
- Logistics: Inventory management and shipment optimization.
- Manufacturing: Assembly and quality control automation.
Industry Applications Matrix
| Industry | Embodied AI Application | Business Benefit |
|---|---|---|
| Healthcare | Patient assistance robots | Improved care efficiency |
| Manufacturing | Autonomous assembly robots | Higher productivity |
| Logistics | Warehouse robots | Faster fulfillment |
| Retail | Smart inventory robots | Reduced stock errors |
| Agriculture | Autonomous harvesting systems | Increased crop yield |
Example
Warehouse robots can autonomously pick, sort, and transport inventory while adapting to changing layouts.
Source: International Federation of Robotics (IFR)
Link: https://ifr.org
by Nabila Altenpi (https://unsplash.com/@itsnabilaaltenp)
Furthermore, education will benefit from personalized learning experiences through AI-driven tutoring systems. These advancements are expected to make learning more engaging and accessible. As embodied AI continues to evolve, it will have an ever-growing effect across all sectors of society, reshaping how we work and live.
The Rise of Autonomous Robots: Everyday Life and Industry
By 2030, autonomous robots will be part of our everyday routines. Self-driving cars will take us to work, while other self-automated household machines will handle daily tasks at home.
In an industrial setting, the transformation will be even more remarkable. These same machines will perform hazardous and repetitive jobs. Factories will also benefit from these machines as they will enhance safety and efficiency by learning and adapting.
The service sector will also see significant improvements in customer experiences through the use of AI. Restaurants may use robots for cooking and cleaning to eliminate human error.
- Transportation: Self-driving cars for safer commutes.
- Home Automation: Robots managing household chores.
- Industry: Machines enhancing safety and productivity.
by Cash Macanaya (https://unsplash.com/@cashmacanaya)
These systems promise to change the way we interact with technology. This embodied intelligence will not only change how we relate to technology, but also how it can help to boost efficiencies and open new frontiers in all kinds of industries
Human-Robot Collaboration and Embodied Intelligence
By 2030, human-robot collaboration will be commonplace throughout all industries. The combination of humans and robots will improve productivity and innovation. Humans and robots working together, using embodied intelligence to perform work with a human touch.
These smart machines can use sensory data as input to make decisions based on their understanding of the environment. As these interactions continue to develop, it is reasonable to assume that such human-like interaction will occur at the workplace level.
Robots and humans can also be complementary. Using AI to create better contextual understanding in robots enables them to provide deeper insights that help organizations make smarter, real-time decisions.
- Enhanced Interactions: Robots interpret and respond to human actions.
- Augmented Teams: Machines and humans working side by side.
- Improved Efficiency: Robots amplify human strengths in various fields.
Furthermore, collaboration between humans and robots will enhance creative solutions and problem-solving across different disciplines.
Human-Robot Collaboration Benefits
| Collaboration Area | Human Strength | Robot Strength |
|---|---|---|
| Decision Making | Judgment | Data Processing |
| Creativity | Innovation | Pattern Recognition |
| Pysical Tasks | Dexterity | Endurance |
| Safety Monitoring | Context Awareness | Continuous Monitoring |
| Repetitive Work | Limited | High Efficiency |
Statistic
Research shows collaborative robotics can increase productivity by up to 30% in manufacturing environments.
Source: Deloitte Insights
Link: https://www2.deloitte.com
Ethical, Social, and Regulatory Challenges
With advances in embodied artificial intelligence (AI) and autonomous robots come ethical questions. The possibility that AI could autonomously make decisions raises concerns about who has the authority to decide. It is therefore imperative that these systems operate within an ethical framework.
With respect to AI’s role in society, there are also numerous social ramifications. For example, if AI robots could perform many jobs currently done by humans, this would likely lead to significant unemployment. Therefore, how society accepts new AI technologies will significantly affect both their adoption rates and success.
Therefore, it is essential that regulatory frameworks be developed to enable governments to properly regulate these technologies. For regulations to effectively govern issues such as data privacy, security, and accountability, governments worldwide must work together to establish consistent international standards and policies.
- Ethical Concerns: Decision-making and autonomy in AI systems.
- Social Implications: Impact on jobs and societal acceptance.
- Regulatory Needs: Data protection and international policies.
In navigating these challenges, there needs to be a balance between coexisting with intelligent machines and humans through 2030.
Major Challenges and Future Solutions for Embodied AI
| Challenge | Impact | Expected Solution by 2030 |
|---|---|---|
| Data Privacy | User trust concerns | Stronger AI regulations |
| Safety Risks | Physical harm potential | Advanced safety systems |
| Bias in AI Models | Unfair decisions | Improved training datasets |
| High Costs | Slow adoption | Affordable AI hardware |
| Workforce Displacement | Job transformation | Reskilling programs |
Key Statistic
According to the World Economic Forum, technological advancements may create 170 million new jobs globally by 2030, while transforming many existing roles.
Source: World Economic Forum Future of Jobs Report
Link: https://www.weforum.org/reports/future-of-jobs-report-2025
The Road Ahead: Predictions and Opportunities for 2030
The future of embodied AI is full of both excitement and uncertainty. We anticipate that by 2030, numerous technological advances will significantly expand AI’s current capabilities. These researchers believe that these systems will likely develop an even greater sense of intuitiveness and blend into society without a hitch.
There are many different paths forward for embodied AI. New machine learning algorithms will enable robots to learn and adjust on their own. Innovations in embodied intelligence may generate novel solutions to various complex global issues. Also, advancements in physical AI will help to increase energy efficiency and contribute to improved sustainability.
The predictions for 2030 support the notion that AI can revolutionize virtually every sector. Whether in health care or transportation, embodied AI has the potential to dramatically reshape how industries operate. When AI combines with other emerging technologies, it can also generate new business models and foster economic growth.
- Breakthroughs: Intuitive AI integration.
- Opportunities: Learning and problem-solving capabilities.
- Sector Transformation: Revolutionizing healthcare, transportation, and industries.
Conclusion: Embracing the Future of Physical AI
By 2030, it is very likely that embodied AI and autonomous robots will be an essential part of our lives. They will be changing how we work, how we live, and how we interact with technology. Collaboration across many different areas of study will be important as we develop systems like these.
To create a better future for ourselves, it is also necessary to address the social and ethical concerns associated with the responsible use of physical AI. This way, we will have created opportunities for innovation and creativity by allowing people to use their imagination and ingenuity to develop new ways of using physical AI, thereby making society more modern.
FAQs
- What is embodied AI, and how is it different from traditional AI?
Embodied AI combines AI with a physical form (like a robot) so it can sense, decide, and act in real-world environments. Traditional AI often operates only in digital spaces, while embodied AI must handle physical constraints and unpredictable conditions. - Which industries will be most impacted by autonomous robots by 2030?
Healthcare, logistics, and manufacturing are expected to see the biggest changes, with robots assisting in surgery, managing warehouses, and improving assembly and quality control. Consumer applications in homes and transportation will also expand. - What technological advances are driving robot capabilities toward 2030?
Improvements in machine learning, sensors, actuators, and natural language processing are making robots more accurate, adaptive, and easier to interact with. Better decision-making and perception will enable more complex autonomous tasks. - Will robots replace human workers by 2030?
Many roles will shift, but the article emphasizes collaboration: robots will increasingly handle repetitive, dangerous, or precision-heavy work while humans focus on oversight, creativity, and complex judgment. Workforce reskilling will be important. - What are the biggest ethical and regulatory challenges?
Key concerns include accountability for autonomous decisions, data privacy and security, job displacement, and the need to ensure safe behavior around people. Stronger regulations and international standards will be needed to guide responsible adoption.
