Synthesia warns speculative AI risks could stifle UK tech sector

British artificial intelligence firm Synthesia has cautioned that exaggerated fears regarding the technology could result in poorly designed regulation, potentially hindering the growth of the UK tech sector. As political pressure mounts for ministers to strengthen the nation’s AI rules, the London-based company argues that the current discourse is driven by speculative scenarios rather than empirical evidence. Alexandru Voica, head of corporate affairs and policy at Synthesia, stated that there is no proof to support claims that AI systems are already out of control. He described the narrative as being propagated by a small group in Silicon Valley who, despite their lack of evidence, employ skilled lobbyists to instil fear in the UK market. Voica characterised these critics as talented science fiction writers whose warnings do not reflect the current reality of the technology.

These remarks follow a series of high-profile warnings from leading figures in the AI industry regarding the risks posed by increasingly powerful models. Dario Amodei, chief executive of Anthropic, has advocated for a slowdown in model development to allow safety measures to catch up, while Sam Altman of OpenAI has supported the case for slowing development where necessary. Jacob Coxon, a former Anthropic researcher, recently resigned from the company, warning that leading laboratories were racing towards self-improving superintelligence and gambling with public safety. Such statements have intensified calls for stricter oversight in Westminster. Parliament’s Joint Committee on Human Rights recently concluded that the UK’s existing regulatory framework is not fit for purpose. The committee recommended the introduction of a new AI Bill and the establishment of an independent statutory regulator, suggesting that high-risk systems should require pre-deployment approval while lower-risk applications face fewer restrictions.

Voica argued that regulations targeting hypothetical catastrophic risks could inadvertently burden routine business uses of AI. He noted that a British company integrating AI models into customer service, software development, or internal workflows could face additional approval processes, liability questions, and technical restrictions. He urged policymakers to focus on making rules work effectively for businesses and consumers, rather than reacting to theoretical risks or media reports from what he termed the Silicon Valley AI doomer bubble. Synthesia, which sells generative AI software to businesses, would itself be subject to any new rules. It is not alone in its opposition to the latest regulatory calls. Aidan Gomez, co-founder and chief executive of Cohere, has also criticised proposals from larger Silicon Valley labs, arguing that the industry needs evidenced standards rather than a cartel. Gomez supported independent testing of powerful AI systems but warned against allowing a handful of dominant developers to dictate industry-wide standards. Cohere advocates for rules based on demonstrated capabilities and risks, alongside mandatory transparency and independent testing for areas such as cyber attacks, fraud, voice cloning, and critical infrastructure.

The Joint Committee on Human Rights highlighted that current UK laws largely apply when AI is deployed, leaving responsibility with users rather than developers. The committee noted that regulators currently lack the power to prevent the release of systems considered unacceptably risky. Committee chair Alex Sobel stated that Britain is unprepared for the potential consequences of AI and called for mandatory transparency requirements and stronger obligations across the AI supply chain. The government has indicated it may introduce further legislation. A government spokesperson told City AM that the risks from advanced AI are taken extremely seriously and that the approach must keep pace with rapidly evolving technology. The spokesperson added that while the existing framework may not always be sufficient, any future measures will be evidence-based and focused on addressing specific risks. Voica also criticised the EU’s approach, noting that businesses still lack clarity over parts of its AI regime despite the legislation having passed several years ago. He compared the current debate to Britain’s nineteenth-century red flag laws, which imposed strict restrictions on early motor vehicles. Voica argued that overly prescriptive rules could similarly hold back a developing British industry and urged governments to work with responsible businesses to understand what works in the real world, thereby aiding economic progress rather than slowing it down.