Who Should Decide How Drones Are Used?

Skydio’s autonomous drones can inspect power lines, search burning buildings and reach an emergency scene before a patrol car. Variants of the same technology can also provide military reconnaissance, track people from the air and potentially contribute to lethal operations.

That range of uses makes Skydio more than a hardware company. It places the Californian manufacturer inside one of the most difficult questions facing the technology sector: when a product can be used by police forces and armed forces, who should decide where the limits lie?

Adam Bry, Skydio’s co-founder and chief executive, has offered an unusually direct answer. Technology companies should not draw their own absolute red lines around legitimate government uses of drones, he argues. Decisions involving military force and public authority should ultimately be made by democratically accountable institutions rather than by a small group of founders and engineers in Silicon Valley.

It is a position that sounds straightforward until the technology moves from theory into the air.

From Flying Cameras To Autonomous Systems

Skydio began with consumer drones capable of following cyclists and avoiding obstacles. It has since withdrawn from the consumer market and concentrated on public safety, infrastructure inspection, national security and defence.

Its current products are better understood as autonomous flying sensors. They can navigate difficult environments, stream live video, avoid obstacles and carry out parts of a mission with limited manual control. Skydio markets these systems to utilities, emergency services, law-enforcement agencies and defence customers.

For a utility company, the attraction is practical. A drone can inspect a transmission line, bridge or industrial site without sending a worker into a hazardous position. For emergency services, it can provide an aerial view of a fire, accident or missing-person search. For police, it can be launched in response to a call and transmit images before officers arrive.

Skydio says its Drone as First Responder systems can be deployed from fixed docks within seconds, giving operators real-time information before personnel enter a potentially dangerous situation. The company has worked with more than 500 public-safety agencies, according to information published when it introduced its DFR Command platform.

These applications illustrate why the argument cannot be reduced to whether drones are good or bad. The same ability to reach a location quickly and observe it from above may help firefighters understand a building collapse or allow a police department to monitor people in a residential neighbourhood.

The technology does not resolve that distinction. Institutions have to.

Silicon Valley’s Red-Line Problem

Parts of the technology industry have previously imposed their own limits on government work. Employees have protested against military contracts, AI companies have published restrictions on weapons applications and some founders have ruled out particular customers or uses.

Those interventions often reflect legitimate concerns. Engineers may not want their work incorporated into lethal weapons. Employees may distrust the institutions buying the technology. Companies may also recognise that complying with the law is not always sufficient to prevent harmful or abusive deployment.

Bry’s objection concerns who ultimately possesses the authority to make these decisions. He argues that a private company should not substitute its judgement for that of elected governments and their institutions, particularly in matters of national defence.

In relation to potentially lethal military uses, his position is that the rules should be established through democratic government, law and military command rather than through unilateral restrictions devised by a supplier. Skydio therefore does not present itself as the final moral arbiter of how an authorised military customer may use its equipment.

There is a defensible logic here. A founder has no electoral mandate, and corporate policies are rarely created through transparent public deliberation. Allowing a few technology executives to determine which national-security policies are legitimate could give private companies extraordinary political power.

Yet democratic authorisation does not remove corporate responsibility. Governments can operate under inadequate legislation, oversight may be weak and procurement decisions can move faster than public understanding. A company still chooses what it develops, which contracts it pursues and what safeguards it builds into its systems.

The real question is therefore not whether Silicon Valley should govern in place of democratic institutions. It is what obligations remain with a company after it has accepted that governments make the final decision.

Public Safety Creates A Different Test

The tension becomes particularly visible when drones are used by police departments.

A Drone as First Responder programme can provide useful information during a burglary, traffic collision, fire or search-and-rescue operation. An aerial view may help dispatchers determine whether officers are required, whether a suspect is armed or whether emergency crews can approach safely. Skydio also argues that earlier information can reduce unnecessary pursuits and dangerous entries.

Residents may nevertheless experience the same programme as a new layer of persistent aerial surveillance. The distinction between responding to a specific emergency and routinely observing public space can become difficult to assess from the ground. Questions arise over when a drone may be launched, what its camera records, how long the footage is retained, who can access it and whether it can be combined with facial recognition or other analytical systems.

These are not speculative concerns. The US Government Accountability Office has found that law-enforcement use of drones and other monitoring technologies in public spaces can create privacy, civil-liberties and bias risks. It has also identified gaps in policies governing how collected information is handled and how risks are assessed.

Bry acknowledges that companies should not decide public-safety policy alone. In April 2026, he argued that Drone as First Responder programmes would succeed only when agencies provided meaningful transparency and communities could see how the technology was being used. He stated that companies should not have the final word over public-safety deployments and placed responsibility with public agencies, communities and democratic processes.

This is a more demanding standard than publishing a general privacy statement. Meaningful accountability would include publicly available operating policies, flight records, defined retention periods, restrictions on unrelated surveillance and independent mechanisms for complaints and review.

A city should be able to explain not only what its drones are technically capable of doing, but what its officers are permitted to do with them.

Regulation Is Moving Towards Scale

The US regulatory environment is also entering a consequential phase.

Many economically valuable drone operations require aircraft to fly beyond the operator’s visual line of sight. Until recently, these flights generally depended on waivers or individual approvals, limiting the ability of companies and public agencies to build large, repeatable programmes.

In August 2025, the Federal Aviation Administration proposed a framework intended to make beyond-visual-line-of-sight operations routine under defined requirements. The proposal covers aircraft standards, operational authorisations, separation from other aircraft, security, record-keeping and responsibility for flights. It could support applications including package delivery, agricultural work, aerial surveying, infrastructure inspection and public-interest operations.

The significance lies less in any single delivery service than in the shift from exceptional flights to normalised operations. Once drones can operate remotely across larger areas without a nearby visual observer, one operator may supervise missions over infrastructure, industrial facilities or urban districts from a central location.

The FAA describes this progression as a gradual move from integrating drones into the national airspace to normalising their presence. Remote ID rules already require most drones to broadcast identifying and location information, improving accountability while supporting more advanced operations.

Greater scale will make governance more important, not less. A pilot project involving occasional flights may be managed through informal safeguards. A permanent network of autonomous aircraft requires clear rules covering airspace safety, cybersecurity, data collection, procurement and public consent.

The Industrial Argument Behind Skydio’s Position

Bry’s argument is also shaped by competition with China.

Chinese manufacturer DJI has long dominated the global civilian drone market. US policymakers have become increasingly concerned about dependence on Chinese-made systems in government, critical-infrastructure and defence applications. Skydio, which manufactures in the United States and has removed Chinese components from its supply chain, stands to benefit from that strategic shift.

Bry argues that restrictions on Chinese technology cannot substitute for building competitive American products. Protection may create an opening, but US manufacturers still need to offer autonomy, reliability, scale and value that customers will choose on performance.

That challenge has become more urgent as the war in Ukraine has demonstrated both the importance and the limitations of small drones. American systems have at times struggled with electronic warfare, GPS disruption, repairability and cost, while Ukrainian forces have continued to rely heavily on adaptable commercial platforms and rapidly modified components.

The experience complicates Silicon Valley’s confidence in software-led defence. Elegant autonomy developed in California must survive jamming, physical damage, limited spare parts and continuous battlefield adaptation. Defence customers are unlikely to accept ethical assurances or advanced specifications in place of operational reliability.

At the same time, investors have poured substantial capital into defence technology. PitchBook data cited by The Guardian indicated that almost $155 billion was invested globally in defence-tech start-ups between 2021 and 2024, compared with $58 billion during the preceding four years. Skydio is part of a wider group of companies seeking to bring faster product cycles and software-defined systems into a sector historically dominated by large contractors.

The debate over red lines is therefore commercial as well as philosophical. Companies that exclude defence work may surrender an increasingly important market. Those that participate gain influence, revenue and government backing, together with responsibility for technologies whose consequences extend beyond conventional enterprise software.

Accountability Cannot Be Outsourced

Bry is right about one important point: decisions about policing, defence and the use of force should not be governed by the personal politics of technology executives.

Corporate red lines can be inconsistent, opaque and reversible. They may change when leadership changes, when investors apply pressure or when a commercially attractive contract appears. Democratic institutions at least offer the possibility of legislation, judicial review, public records and electoral consequences.

But government authorisation cannot serve as a complete ethical exemption for suppliers.

Drone manufacturers decide which capabilities to pursue, how easily systems can be repurposed, what information they record and whether safeguards are technically enforceable. They also control what prospective customers are told about accuracy, autonomy, security and operational limits.

A credible division of responsibility would therefore be shared. Legislators should determine the lawful boundaries. Public agencies should publish operational policies and answer to the communities they serve. Courts and independent bodies should examine abuses. Companies should test their systems rigorously, disclose material limitations and build products that allow customers to be audited.

The most consequential drone debate is no longer whether the aircraft should be allowed to fly. It is how power will be distributed once autonomous systems become a routine part of public infrastructure, policing and defence.

Silicon Valley should not make those decisions alone. It should not be permitted to escape them either.