Surveillance on Wheels: Why the Government's Fast-Track to Autonomous Vehicles Threatens Our Privacy, Democracy and Workers’ Rights

The Starmer government has fast-tracked autonomous vehicle deployments to spring 2026, a full year ahead of schedule. Major tech companies including Waymo (Google), Baidu, and Tesla are preparing to flood London’s streets with sensor-equipped vehicles that can track, record, and monitor every aspect of public life. Yet there has been virtually no public discussion about the profound implications of this technological transformation.

The Surveillance Capabilities Are Staggering

Every autonomous vehicle is a mobile surveillance platform carrying:

• 360-degree camera arrays with potential to be used in facial recognition, gait analysis, license plate scanning, and continuous recording of public activities, protests, and social gatherings

• LiDAR mapping systems that create detailed 3D maps of communities, tracking pedestrian movement patterns and building occupancy

• Advanced sensors including radar and infrared that can detect human presence and movement even in low visibility

• Audio recording systems that can capture conversations, political speech, and community interactions

• Wireless monitoring that can track mobile devices and map social networks through proximity analysis
  

Unlike fixed CCTV, these vehicles move continuously through our communities, creating blanket surveillance coverage that eliminates anonymity in public space. In San Francisco, police training documents already instruct officers that autonomous vehicles are “recording their surroundings continuously” and useful for “investigative leads.” Reports in TechSpot, TechCrunch, and Wired show how police have already tapped robo-taxis for mass surveillance — monitoring protests, investigating unrelated crimes. UK deployments risk normalising this same infrastructure.

What the government’s consultation actually says about data and information sharing

To speed up these developments, the government just ran a consultation on a national Automated Passenger Services (APS) permitting regime, one of the biggest changes to transport and public life in decades. The APS consultation proposes a regulatory framework for permits, including “disclosure and use of information” provisions. The proposals provide broad discretionary powers for information sharing without AV-specific privacy safeguards. They:

• Lack explicit prohibitions on surveillance technologies like facial recognition or emotion tracking

• Provide broad permissions for police access to AV data streams

• Rely heavily on existing data protection law, which has proven inadequate in the platform economy

• Include almost no mechanisms for community oversight or democratic accountability
  

Put plainly, the consultation grants wide leeway to share information but doesn’t build AV-specific guardrails to stop over-collection, over-retention, or function creep. That balance needs to be fixed before AV fleets scale.

We’ve seen this direction of travel before

Worker Info Exchange’s research shows how platform data pipelines expand once they exist. In Managed by Bots, we documented how ride-hail platforms became attractive sources of intelligence for UK policing. During Uber’s 2020 licensing appeal, Uber’s UK/Western Europe GM Jamie Heywood described a “maturing relationship” with police bodies, from the National Counter Terrorism Policing Network to the NPCC, and highlighted over 2,000 annual data requests from the Metropolitan Police alone, far outstripping some national totals reported elsewhere. The report also notes the NPCC lobbying TfL in support of Uber’s licence, signalling the growing strategic value of platform data to policing. In a separate case study, a driver was suspended for seven weeks following soft intelligence, not a criminal charge. He was never a suspect, but the lack of due process caused serious reputational and financial harm.

Beyond policing, platform data pipelines are now being enlisted for immigration and tax enforcement. The Home Office recently announced cooperation with major delivery platforms (e.g., Deliveroo, Uber Eats, Just Eat) to intensify identity checks and enforcement against illegal working. Meanwhile, digital platforms are required to report certain user earnings to HMRC. These policies illustrate a common trajectory: once data flows exist, they tend to expand across government uses, often without accessible redress for workers or passengers.

The political context: who gets to build and run this infrastructure?

This debate is not abstract. It is about who we allow to run pervasive sensors in our streets. Some firms bidding for UK AV markets are at the centre of intense political controversy. Recent coverage of Elon Musk’s remote appearance at a far-right rally in London—including statements like “violence is coming” and calls for the dissolution of government—raises serious questions about entrusting surveillance-capable fleets to corporate leaders engaged in inflammatory politics.

What our research shows and why it matters for AVs

Across years of work with drivers and communities, WIE has shown how opaque algorithms and data asymmetries harm the public interest. In Dying for Data, we documented the UK’s public-data deficit: authorities and communities lack the anonymised journey data needed to manage utilisation, congestion, and accessibility, while platforms keep key data locked up. AVs could amplify these harms if oversight replicates today’s black-box arrangements. Conversely, with privacy-preserving open data, communities could actually audit impacts (e.g. fleets of empty vehicles circling city streets, areas left without adequate service, and unpredictable or unfair pricing) without exposing personal information.

The worker disruption no one is planning for

Beyond the risks to privacy and surveillance, autonomous vehicles also pose a direct threat to livelihoods. The government projects 38,000 new jobs by 2035, but this pales against the 381,000 licensed taxi and private hire drivers in England alone whose jobs are threatened by robo-taxi services now being trialled. Literature on automation estimates that 1.3 to 2.3 million driving and transport roles are vulnerable to replacement by autonomous or semi-autonomous systems. Such disruption will disproportionately affect working-class communities and ethnic minorities, who are overrepresented in transport work.

This is not a smooth “transition to better jobs.” Without enforceable protections, displacement is likely to precede any meaningful job creation.

The emerging shadow workforce: deskilling, precarity, and data work

Even as human driving is phased out, AV systems will still require human support — but in more precarious, deskilled forms:

• Remote operators and safety monitors: These roles may be outsourced or subcontracted, with scheduling and performance management controlled via opaque algorithms, leaving workers vulnerable to arbitrary decisions.

• Field agents, cleaners, battery/charger technicians: These are often subcontracted functions, required to respond during antisocial hours, under minimal protections and pay.

• Roadside recovery teams: Handling AV failures or system edge cases under pressure, with risks shifted from operators to often contingent workers.
  

In the gig economy widely, similar dynamics are already seen: platform workers experience algorithmic control, limited recourse, and employment insecurity. The AV context will intensify these risks, layering safety-critical responsibilities onto already insecure work.

What must change

If AVs are to be introduced onto UK streets, robust safeguards are needed to prevent them from becoming roving surveillance platforms. Worker Info Exchange has set out five key requirements in our response to the consultation:

1. Strict limits on surveillance and police access. Operators must publish clear specifications of what data is captured, how it is handled, and when it is deleted. Biometric identification and continuous recording of public space should be prohibited. Police access should be case-specific, warrant-based, and transparently reported.
   

2. Privacy-preserving open data. Operators should be required to publish anonymised, aggregate journey data (e.g., passenger vs empty miles, wait times, occupancy) using proven privacy safeguards. This would enable communities and regulators to assess impacts on congestion, accessibility, and pricing, without exposing individuals.
   

3. Algorithmic accountability. Companies should disclose how vehicle safety systems, dispatch, and pricing algorithms operate; commit to bias testing; and submit to independent audit.
   

4. Utilisation and capacity controls. Permits should include utilisation targets and caps to prevent oversupply, “ghost cruising,” and induced congestion. Capacity limits should be tied to demonstrated demand and integrated with public transport strategies.
   

5. Operator fees and a technology dividend. Firms should contribute through statutory levies on bookings or mileage to fund independent audits, community impact projects, and worker transition support.
   

6. Worker Protections and Just Transition. The framework must require regular reporting on displacement and labour impacts, enforceable work standards across all AV-related roles and structured consultation with unions and worker representatives.

Final word

Autonomous passenger services represent a profound technological transition that will reshape not only transport but also urban surveillance, labour markets, and public accountability. The government’s fast-tracked timetable makes it urgent that the UK embeds safeguards from the outset. Without them, AV deployment risks entrenching a pervasive surveillance infrastructure operated by private firms with limited democratic oversight, while displacing hundreds of thousands of workers into precarious and low-paid roles.