Three Ways to Better Govern the Use of AI

The development of artificial intelligence (AI) offers the potential for significant positive benefits for society. However, there are profound concerns about the potential negative impacts on democracy, fundamental rights, and people’s lives if not governed correctly. Open government and its guiding principles of transparencyAccording to OGP’s Articles of Governance, transparency occurs when “government-held information (including on activities and decisions) is open, comprehensive, timely, freely available to the pub... More, accountability, inclusionOGP participating governments are working to create governments that truly serve all people. Commitments in this area may address persons with disabilities, women and girls, lesbian, gay, bisexual, tr... and participation have a key role to play in ensuring AI is used responsibly and in the service of the public.

OGP and the Government of Kenya convened reformers in government and civil society from 12 countries to discuss ways to govern new and emerging digital technologies. Here are three key conclusions for the governance of AI.

Society Must Define AI

The trajectory of AI needs to be shaped by the global, national, and local challenges we face and the interests of everyone in society – not the private interests of big tech companies and their shareholders. 

This means asking questions of what, why and who: What is the societal challenge that needs to be overcome? How can AI help to address the challenge? And who stands to benefit and lose from its use? Identifying the problem and evaluating how a technology can solve it, and with what consequences, should be integral to conceptualizing, creating, governing, and evaluating the use of AI. 

These principles can be realized in practice by:

• Defining the “public interest” for AI through open, inclusive, and multi-stakeholder conversations, including civil society and the public. This can ensure that AI is developed in a way that is consistent with public values, interests, and concerns, contributing to greater public impact and stronger public trust.
• Maintaining human oversight and accountability for AI and automated decision-makingAs governments are increasingly turning to algorithms to support decision-making for public services, transparency and accountability of algorithm use is necessary to minimize their harm and increase .... Decisions made by an algorithm must be able to be challenged by humans to ensure that their results match their intended purpose. This includes not only having a “human in the room,” but also through pre-deployment testing, ongoing monitoring, and transparent impact and human rightsAn essential part of open government includes protecting the sacred freedoms and rights of all citizens, including the most vulnerable groups, and holding those who violate human rights accountable. T... assessments.

AI must not replicate existing inequalities

The ability to secure the benefits and avoid the risks of AI is not equally shared. The benefits from AI and other digital technologies for humanity can only be fully realized if they are developed with the needs, interests, languages, and contexts of different countries and diverse communities in mind. 

Specifically: 

• Harmful incidents targeting women, youthRecognizing that investing in youth means investing in a better future, OGP participating governments are creating meaningful opportunities for youth to participate in government processes. Technical ..., minorities and other vulnerable groups have demonstrated the importance of procuring, designing, testing, and implementing AI technologies through a human rights and genderOGP participating governments are bringing gender perspectives to popular policy areas, ensuring diversity in participatory processes, and specifically targeting gender gaps in policies to address gov... lens. 
• Digital technologies are not neutral, but instead reproduce biases inherent in their data, programming and uses. Truly tackling these biases in algorithms requires addressing biases in our institutions and societies, but robust public complaints mechanisms, regular independent auditsInstitutional and legal frameworks are necessary for providing assurance of the integrity of financial information and of compliance with budgetary rules and procedure. Technical specifications: These..., and strong regulatory enforcement can help to root out and address these biases in AI.

 Global norms are important but mean little without country reforms  

While global governance is important, it needs to be translated into meaningful country-owned regulations and reforms to ensure the benefits of digital technologies to society are maximized while minimizing risk. National reformers from governments and civil society are key to leading this agenda.

Open Gov reformers can lead the way by:

• Developing a strong national policy framework for AI, through an open and inclusive process, that provides assurances to the public that risks are managed and responsible use of AI is possible. Robust governance creates the context for more innovation, not less. Conversely, the lack of a framework heightens perceived and actual risks and may lead to public backlash and the abandonment of potentially valuable solutions. 
• Using open, inclusive and accountable procurement processes to ensure oversight and robust decision-making, even in the absence of a strong legislative framework.

Open government values of transparency, participation, inclusion, and accountability are key to ensuring that the adoption of digital technologies, such as AI, helps to improve the lives of people and communities and strengthen democracies. Check out the Open Gov Guide on digital governance for more recommendations and examples.