AI intimacy bans and drone noise: The UK innovation dilemmas

The line between technology and human emotion is being redrawn in Canberra, while London’s skies are getting quieter—if only slightly. Two developments this week highlight how innovation is increasingly shaped not by what engineers can build, but by what societies will tolerate.

Australia’s digital duty of care: When AI becomes too close

Australia is poised to become the first country to explicitly ban AI chatbots designed to foster “emotional intimacy.” The proposed Digital Duty of Care laws, which passed the lower house this week, would prohibit platforms from deploying AI agents that simulate friendship, romantic attachment, or emotional dependency. The move follows Prime Minister Anthony Albanese’s warning that children are treating these bots as “best friends,” raising concerns about psychological harm and social isolation.

The legislation reflects a broader shift in how governments approach AI regulation. Unlike the UK’s focus on security and infrastructure—evidenced by last week’s £700m AI security pledge—Australia is targeting the behavioural effects of technology. The bill also includes provisions to expand the powers of the Telecommunications Industry Ombudsman, allowing it to investigate complaints against social media and AI platforms and publicly name companies found in breach.

Critics argue the law risks overreach, stifling therapeutic applications of AI, such as mental health companions for elderly users. But supporters counter that emotional manipulation by profit-driven algorithms demands urgent intervention. The debate mirrors tensions in the UK, where AI datacentres have faced protests over energy consumption, but where emotional AI remains largely unregulated.

For British policymakers, Australia’s approach presents a dilemma: should the UK adopt similar safeguards, or does its emphasis on AI as an economic driver preclude such restrictions?

The sound of progress: Why drones are getting a muffler

As drone deliveries edge closer to becoming an everyday reality in the UK, manufacturers are confronting an unexpected challenge: noise. A BBC investigation this week revealed that companies like Amazon, Wing, and Zipline are racing to develop quieter propellers, electric motors, and even “whisper modes” to reduce the high-pitched buzz that has sparked complaints from communities beneath flight paths.

The issue is more than a nuisance. Noise pollution is emerging as a key regulatory hurdle for drone expansion. In the US, the Federal Aviation Administration (FAA) has already imposed strict noise limits on commercial drone operations, while the EU is developing its own standards. The UK’s Civil Aviation Authority (CAA) has yet to set formal rules, but local councils are beginning to push back. In Milton Keynes, where drone trials have been underway since 2022, residents have reported sleep disruption and stress, with some describing the sound as “like a swarm of bees.”

The industry’s response has been twofold. First, redesigning propellers to minimise turbulence—inspired by owl feathers, which allow silent flight. Second, adjusting flight paths to avoid densely populated areas, even if it means longer delivery times. Wing, a subsidiary of Alphabet, claims its latest drones are 30% quieter than models from two years ago, while Amazon’s Prime Air team is testing “acoustic camouflage” techniques to blend drone noise into urban soundscapes.

Yet the challenge remains: can drones ever be quiet enough for widespread acceptance? Unlike electric cars, which benefited from decades of incremental noise reduction, drones are being introduced at scale with little public consultation. The CAA’s forthcoming Drone Noise Framework, expected next year, will be a critical test of whether the UK prioritises convenience over quality of life.

The quiet crisis in AI-driven policymaking

While Australia’s chatbot ban and the UK’s drone noise debate dominate headlines, a less visible but equally significant trend is unfolding: the rise of AI-generated public consultation responses. Research published this week in The Guardian revealed that three-quarters of submissions to the UK government’s 2023 consultation on banning disposable vapes were produced by AI bots. The campaign, orchestrated by pro-vaping lobby groups, flooded the process with identical, machine-generated arguments, marking what experts believe is the first large-scale attempt to manipulate health policymaking using AI.

The incident exposes a growing vulnerability in democratic processes. Public consultations, designed to gauge citizen sentiment, are increasingly susceptible to automated astroturfing—where bots create the illusion of grassroots support or opposition. The UK’s Department of Health dismissed the fake responses, but the episode raises questions about how future consultations will be safeguarded. Should governments require digital verification for submissions? Or will AI’s ability to mimic human input render the entire process obsolete?

The implications extend beyond vaping. As AI tools become more accessible, industries facing regulation—from fossil fuels to social media—could deploy similar tactics to sway policy. The UK’s Competition and Markets Authority (CMA) is already investigating whether AI-generated content could distort competition, but no agency has yet proposed concrete solutions.

For now, the vaping consultation stands as a cautionary tale: in the race to regulate AI, governments may find themselves outmanoeuvred by the very technology they seek to control.

What to watch

1. Australia’s Senate vote: The Digital Duty of Care bill faces its next hurdle in November. If passed, it could set a global precedent for AI emotional regulation.
2. CAA’s drone noise rules: Expected in early 2027, these will determine whether the UK adopts a permissive or precautionary approach to drone expansion.
3. AI consultation safeguards: The UK government has yet to announce measures to prevent future bot-driven manipulation of public consultations.

Innovation is no longer just about what’s possible—it’s about what’s permissible, and what’s bearable. This week’s stories suggest that the next frontier isn’t in labs or server farms, but in the messy, human spaces where technology meets society.