Primer: Introducing a “Hippocratic Oath” and licensing regime for engineers and research scientists working on Frontier AI

The UK Government should commission a white paper on the viability of a professional licensing system for engineers and research scientists that work on the technical development of Frontier Artificial Intelligence.

As part of ongoing AI cooperation, the UK Government should propose a working group with the US and EU on the establishment of a mutual licence system for engineers and research scientists, working on the technical development of Frontier AI. This system would complement ongoing international efforts to coordinate and cooperate on AI regulation.

• Individual engineers and research scientists currently working on the technical development of Frontier AI have a far more profound impact on members of the public than a doctor or a lawyer, and an equal social responsibility to uphold ethical standards, yet as a profession it remains largely unregulated.

• Like doctors, there would be inherent value for society for engineers and research scientists working on Frontier AI to be led by a Hippocratic oath that the development of emerging technology should be underpinned by the principle of “doing no harm”.

• This could ensure that engineers and research scientists working on technology consider the case for embedding democratic principles and ensuring their work is not used to spread disinformation.

Incentivising the tech sector

• As part of the white paper, Ministers should consult the views of tech companies of different sizes on the viability of a professional licensing scheme, as well as consider discussions about incentivising adoption.

• Incentives for tech companies who are willing to support the roll-out of a licensing scheme could include tax incentives, increased global talent visas, and the exploration of government contracts for public data access.

• In the past tech companies have made voluntary pledges to mitigate the harm of AI, this includes the pledge at the 2024 Munich Security Conference signed by 25 tech companies to work together to detect and counter harmful AI content in elections. A similar voluntary pledge could be proposed for tech companies to support the introduction of a licensing scheme.

HOW IT WOULD WORK IN PRACTICE?

There are several models of how this could operate:

Model One could see the licensing system emulate the professional standards bodies in the legal, health, and accountancy sectors, where professional standards bodies keep a list of those who are licenced, oversee the granting of licences, and any infraction is dealt with by independent standards bodies.

Model Two could see the Government create a direct licensing system and regulatory body to oversee its enactment and enforcement. As a first step this would require technology companies to disclose to the Government those individuals working on Frontier AI development. As mentioned above this group should include engineers and research scientists working on the technical development of new Frontier AI models.

Model Three could see the technology companies themselves create a licensing system for engineers working on Frontier AI which in turn is overseen by a government-run enforcement body. This might take the form of a membership organisation that is self-regulating rather than on an individual basis.

Similarly, there is no reason why a government could not consider a policy solution that takes from all three models.

WHAT ARE THE BENEFITS?

Transparency, a licensing system, would bring significant transparency for the government and the public regarding the individuals who are currently developing Frontier AI that will define the next few eras of geopolitical, globalisation, democracy, and employment.

Accountability, a licensing system, would offer a positive reinforcement mechanism whereby individual engineers are accountable for the technology that they produce and therefore consider ethical guardrails out of self-interest. For tech companies it would also offer clarification when it comes to legal liabilities over any potential future cases regarding the deployment of Frontier AI.

Compliance, a licensing system, would allow technology companies, professional bodies, and regulators to identify bad actors in the space who are seeking to utilise technology for harm.  

Raising Professional Standards, a licensing system could be a part of a broader drive to raise ethical standards across the technology sector with tech companies having a positive incentive to hire engineers and research scientists for Frontier AI projects who are licensed and licensed engineers and research scientists are able to command higher pay, better work conditions, and would have cover to push back against any corporate demand to develop AI in an unethical way. 

Attraction of Talent, a licensing system would make it easier for tech companies to work with the government to recruit global talent to work on Frontier AI and offer a further layer of protection when it comes to recruitment of staff from countries who are deemed AI rivals to the UK. In turn it could pave the way for a special migration category for those working on Frontier AI which adds benefits to the economy.

Cooperation, a licensing system, would be an area of AI cooperation for the UK with key partners and competitors including the PRC. A shared-list of engineers and research scientists working on Frontier AI development might be a first-step towards the kind of regular communication nuclear scientists had throughout the Cold War to avert a nuclear disaster.

AN IMPORTANT PILLAR OF REGULATION, NOT A PANACEA 

• The adoption and cooperation on a shared licensing system for engineers and research scientists working on frontier AI, should not be viewed as a panacea in itself but rather as an important pillar of ongoing AI regulation. 

• For AI regulation to be effective, there needs to be sensible and robust regulations for nation states, private sector companies, and individual engineers and research scientists designing, deploying, and using AI. Ensuring the balance between supporting innovation and economic growth, while protecting the public and ensuring democratic resilience.

Cover Image Courtesy of rawpixel.com