- with Inhouse Counsel
- with readers working within the Pharmaceuticals & BioTech industries
Over the past year, we sat down with four voices offering complementary vantage points on robotics: Jason Fiorillo (Chief Legal Officer, Boston Dynamics), Magnus Egerstedt (Dean of Engineering, UC Irvine), Joyce Sidopoulos (Co-Founder and Chief of Operations, MassRobotics), and Tyler Cook (Assistant Program Director, Emory Center for AI Learning; Professional Fellow, Emory Center for Ethics; formerly Georgia Tech). In candid conversations led by John Lanza and Jen Audeh, they reflected on regulation and safety, data and security, workforce dynamics, adoption hurdles, IP and AI-assisted invention, and where the field is headed. What follows synthesizes those interviews into the top themes we heard-using the interviewees' own framing and emphases-and ties together their overlapping views and points of tension.
Theme 1: Safety, Regulation, and Accountability Are Playing Catch-Up with Mobile, Learning Robots
Across the conversations, there was clear agreement that today's safety regimes were built for a different era. As Jason Fiorillo put it, traditional frameworks assumed fixed, caged machines; but once robots move through facilities and operate with less direct supervision, the old "big red button" model shows its limits. He expects standards to evolve as robots become more physically capable and enter settings that never housed robots before, with thorny questions about split-second decisions and how to assign responsibility among manufacturers, programmers, operators, and autonomous systems. Magnus Egerstedt echoed that we are not at a point where society should "bet our lives" on AI, emphasizing that new standards must address both physical safety and the privacy implications of robots that learn in homes, hospitals, and public spaces. Both stressed that regulation has to keep pace with autonomy and learning, not just mechanical risk. Looking ahead, they anticipate new standards, updated regulations, and even novel insurance or risk-sharing models to sort out accountability when autonomous behavior leads to harm.
Theme 2: Data Governance and Cybersecurity by Design Are Prerequisites for Trust
The interviews converged on the same baseline: robots are "mobile sensor platforms," and that raises the stakes for privacy and security. Jason Fiorillo highlighted how routine operations can sweep up sensitive information-images, conversations, and other data that go well beyond familiar privacy frameworks-making it imperative to think holistically about safety, privacy, and autonomy rather than in isolation. Magnus Egerstedt stressed building "assured autonomy" and safe learning into systems from the outset, not as an afterthought, with practices like adversarial testing and security-first architectures. Tyler Cook emphasized the ethical dimension of data use: transparency about collection and sharing, encryption and other safeguards, and ongoing mitigation of AI bias in training and deployment. For all four, trust will come from concrete security engineering, clear disclosures, and continuous monitoring-not just policy statements.
Theme 3: Workforce Change Is Here—Upskilling and Human–Machine Teaming Are the Levers
On workforce impacts, the group largely rejected a simplistic "robots replace workers" narrative. Joyce Sidopoulos sees persistent labor shortages across manufacturing, agriculture, construction, and healthcare, with robots taking on "dull, dirty, and dangerous" work while people set up, program, and maintain systems. Her view is that acceptance is growing when deployments are participatory and when workers are offered real retraining and growth paths. Magnus Egerstedt pointed to the premium on lifelong learning and adaptability, arguing that universities and employers alike need to rethink education for human-machine teaming. Jason Fiorillo noted that near-term shifts may be more pronounced in white-collar roles as AI tools reshape productivity, while robotics will continue to address structural gaps in manual labor without an immediate, sweeping displacement. Tyler Cook added a caution: as automation advances, leaders should remain mindful of deskilling risks and preserve human autonomy in decision loops.
Theme 4: Adoption Barriers Are About ROI, Integration, and Change Management—not Technical Possibility
When it comes to getting robots into real workflows, the hardest problems are rarely purely technical. Joyce Sidopoulos framed the central questions as business-model issues: upfront capital, uncertain ROI, limited in-house expertise to implement and maintain systems, and uneven standards coverage in newer use cases. She shared practical playbooks that reduce resistance-engage frontline workers to identify least-liked tasks, run structured retraining, and even humanize systems (naming robots, celebrating milestones) to cultivate ownership. Magnus Egerstedt underscored that organizations should start with narrow, well-bounded problems to prove value and safety before scaling. Across the interviews, there was support for public-private mechanisms-financing, grants, standards work, and ecosystem hubs-to help startups bridge the gap from research to commercialization and help adopters de-risk integration.
Theme 5: IP Norms Are Shifting as AI-Assisted Invention Becomes Ubiquitous, but Openness Still Matters
The interviews captured a pragmatic balance between openness and protection. Magnus Egerstedt expects a continued mix: open platforms for broad experimentation alongside proprietary technologies that fund commercialization, with success depending on striking the right balance. Jason Fiorillo focused on a fast-emerging frontier: as AI tools become inseparable from R&D, the line between human invention and machine assistance will keep pressure on long-standing patent concepts. He noted that AI-assisted inventions are not categorically excluded today, but practical questions remain about attribution, disclosure, and what it means for an invention to be "made by man" when code and models contribute materially. Tyler Cook added that these shifts amplify the need for explainability and responsible use in both invention and deployment.
Cross-Currents to Watch: Sustainability and Supply Chains
Two threads cut across the discussions. First, sustainability: Magnus Egerstedt flagged the energy cost of compute for advanced autonomy and called for efficiency gains in algorithms and hardware, as well as life-cycle thinking from manufacturing through disposal. Second, supply chains: Jason Fiorillo described a more complex geopolitical landscape, urging diversification of suppliers, inventory buffers for critical components, and a tilt toward allied sourcing-practical steps even if they cannot eliminate risk entirely.
What to Expect Next
Over the next stretch, expect robots to keep moving out of fenced industrial cells and into human-proximate settings, with safety, privacy, and accountability standards evolving in parallel. As Fiorillo, Egerstedt, Sidopoulos, and Cook each stressed, near-term progress will come from pragmatic deployments that prove value and safety, stronger data governance and cybersecurity by design, deliberate upskilling and change management, and a practical balance between openness and protection in IP as AI-assisted invention becomes routine. We will continue the conversation with the industry through additional interviews, roundtables, and updates as the landscape shifts, and we will report back on what we are hearing and how it is shaping best practices.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
