Our current operations, products, and services use artificial intelligence (“AI”), including machine learning. Examples of our current uses of machine learning include (i) using algorithms to process video and system data to identify surgical activities and surgical performance indicators to support learning, teaching, and practice management, and (ii) using algorithms to support surgical planning and navigation. Future innovations in our products and services will likely continue to incorporate AI, and these applications may become important in our operations over time, for example, our development of machine learning-enabled medical devices (“MLMDs”).
As with many technological innovations, there are significant risks and challenges involved in maintaining and deploying these technologies, and there can be no assurance that the usage of such technologies will enhance our products or services or be beneficial to our business, including our efficiency or profitability.
Our ability to continue to maintain or use such technologies may be dependent on access to specific third-party software and infrastructure, such as processing hardware, and we cannot control the availability or pricing of such third-party software and infrastructure, especially in a highly competitive environment. Our products and services may not compete effectively with alternative products and services if we are not able to source and integrate the latest technologies into our products and services. In addition, a number of aspects of intellectual property protection in the field of AI are currently under development, and there is uncertainty and ongoing litigation in different jurisdictions as to the degree and extent of protection warranted for AI technologies and relevant system input and outputs. If we fail to obtain protection for the intellectual property rights concerning our AI technologies, or later have our intellectual property rights invalidated or otherwise diminished, our competitors may be able to take advantage of our research and development efforts to develop competing products, which could adversely affect our business, reputation, financial condition, or results of operations. Refer to our risk factor titled “If we are unable to fully protect and successfully defend our intellectual property from use by third parties, our ability to compete in the market may be harmed” for additional risks related to intellectual property.
The regulatory landscape surrounding AI is also evolving, and the use of machine learning technologies may expose us to an increased risk of regulatory enforcement and litigation. For example, in October 2023, the FDA, Health Canada, and the U.K.’s Medicines and Healthcare products Regulatory Agency jointly published the “Predetermined Change Control Plans for Machine Learning-Enabled Medical Devices: Guiding Principles.” These principles may require significant regulatory oversight, such as additional premarket review, and ongoing regulation through monitoring, maintenance, and improving device performance to ensure safety and effectiveness.
In the U.S., an executive order was issued in October 2023 on the Safe, Secure and Trustworthy Development and Use of AI, emphasizing the need for transparency, accountability, and fairness in the development and use of AI, including in the healthcare industry. The order seeks to balance fostering innovation with addressing risks associated with AI by providing eight guiding principles and priorities, such as ensuring that consumers are protected from fraud, discrimination, and privacy risks related to AI. The order also calls for future regulations from various agencies, such as the Department of Commerce (to draft guidance for detecting and authenticating AI content) and the Federal Trade Commission (to ensure fair competition and reduce consumer harm). In alignment with the order, other agencies have published guidance.
Apart from the U.S., policymakers in key jurisdictions, such as the EU, are actively working on legislation and regulations to encourage the development and use of ethical and safe AI technologies. For example, on April 21, 2021, the European Commission proposed a regulation seeking to establish a comprehensive, risk-based governance framework for AI in the EU market (“EU AI Act”). The proposal is intended to apply to companies that develop, use, and/or provide AI in the EU and includes requirements around transparency, conformity assessments and monitoring, risk assessments, human oversight, security, and accuracy. In addition, on September 28, 2022, the European Commission proposed two Directives seeking to establish a harmonized civil liability regime for AI in the EU. These regulatory proposals are at varying stages of the legislative process and are not yet finalized; the EU AI Act is at an advanced stage and the text is currently expected to be finalized by the end of 2023. Once finalized and in force, this regulatory framework is expected to have a material impact on the way that AI is regulated in the EU and, together with developing guidance and/or decisions in this area, may affect our use of AI and our ability to provide, improve, or commercialize our services, require additional compliance measures and changes to our operations and processes, result in increased compliance costs and potential increases in civil claims against us, and could adversely affect our business, operations, and financial condition.
Other jurisdictions where we operate have already or are also expected to introduce guidelines and regulations around the use of AI within the next few years. The regulations may impose onerous obligations and may require us to rework or reevaluate improvements to be compliant, potentially increasing costs.
Moreover, some of the AI features of our products involve, or may involve, the processing of personal data and may be subject to laws, policies, legal obligations, and codes of conduct related to privacy and data protection, each of which may be interpreted in ways that may affect the way in which we engage with machine learning and require us to make changes to our business practices and products to comply with such obligations. Our use of AI technologies may involve the storage and transmission of confidential or sensitive information, including personal information of employees, customers, and others, as well as protected health information of clients’ patients. In addition, due to the sensitive nature of the information, the security features of our computers and systems, network, and communications systems infrastructure are critical to the success of our business. A breach or failure in our security measures could occur from a variety of circumstances and events, including third-party action, employee negligence or error, malfeasance, computer viruses, cyber-attacks, or ransom-related attacks by computer hackers, failures during the process of upgrading or replacing software and databases, power outages, hardware failures, telecommunication failures, user errors, or catastrophic events, and any of the foregoing events could have a material adverse effect on our business, financial condition, or results of operations. For more information on risks associated with the processing of confidential and sensitive information, including personal information, refer to our risk factor titled “Information technology system failures, cyberattacks, or deficiencies in our cybersecurity could harm our business, customer relations, financial condition, or results of operations.”
Though we have taken steps to be thoughtful in our development, training, and implementation of machine learning, including taking steps to comply with the laws and frameworks discussed above that are currently in effect, our machine learning-related processing could pose certain risks to our customers, including patients, clinicians, and healthcare institutions, and it is not guaranteed that regulators will agree with our approach to limiting these risks or to our compliance more generally. Risks can include, but are not limited to, the potential for errors or inaccuracies in the algorithms or models used by the MLMDs, the potential for bias or inaccuracies in the data used to train the MLMDs, the potential for improper processing of personal information that could lead to deprecation of our algorithms, and the potential for cybersecurity breaches that could compromise patient data or device functionality. Such risks could negatively affect the performance of our products, services, and business, as well as our reputation and the reputations of our customers, and we could incur liability through the violation of laws or contracts to which we are a party or civil claims.