Generative artificial intelligence’s (AI) pace of innovation and scale of deployment, combined with its novel potential and risks, is an urgent call to take meaningful action.
The time to act is now: By most accounts, we have only two to three years before the models will become too sophisticated and the technology too embedded to track, manage, audit, or inspect. At this moment, issues are still emergent, and society has an opportunity to shape its future.
“Rarely has a transformative technology with such high stakes, societal implications presented itself anew.”
Philanthropy—alongside civil society, policymakers, and industry—has a clear opportunity to help bend the course of generative AI toward its large-scale, potential benefits like accelerating drug discovery, addressing the climate crisis, and augmenting human capacity. And as leaders, we have an imperative to steer generative AI away from both current and extreme harms, which include bias and discrimination in automated decisions, sophisticated mis- and disinformation, the elimination of occupations and industries, and bioweapons.
To do so, we must look at the digital technology system as a whole and everyone who is affected, rather than in stages or piecemeal. We must focus on a technological future that will serve society for years to come.
Omidyar Network has engaged hundreds of experts and practitioners: generative AI executives and chief scientists; venture capitalists and investors; lawmakers and policymakers; researchers and activists; our fellow philanthropists; artists and academics. Nearly all have underscored the hyper-normal pace at which this technology is evolving and the multiple paths it might take. A lot remains unknown, and even experts disagree about which priorities require urgent attention. But there is little disagreement that attention is merited. As such, we approach this task with curiosity, humility, and creativity. Even as we plan to act quickly, we know there will be surprises. Urgency must be matched with flexibility and adaptability.
Omidyar Network’s position on generative AI reflects our perspective on digital technology as a whole: generative AI should benefit society, rather than individuals and communities having to adapt to accommodate it. Nested within this broad point of view, we have three deeply held principles relevant to generative AI, which will guide our strategy:
We believe in inclusive and meaningful participation by different stakeholders, actively working together to craft what this technology can be so it can best serve society, as well as what it must never become. Omidyar Network is explicitly prioritizing who the tech is made for and by whom. We believe that technology’s power and decision making should not be concentrated in the hands of a few, whether corporations, governments, or the wealthy. This transformation will only happen when individuals with diverse lived experiences and perspectives actively contribute to the design, deployment, monitoring, and effects of generative AI. This will require building broad-based coalitions that work together toward collective good.
While much remains unknown, we can posit this technology will have an impact on our social fabric that is already frayed due to bias, discrimination, and disinformation. We are concerned about the impact generative AI could have on people’s sense of belonging, potentially exacerbating and weaponizing feelings of loneliness and isolation, which the US Surgeon General has identified as a growing epidemic. We also anticipate the harms will be felt most acutely by communities of color, emerging economies, and young people that have been intentionally and systemically left out in the past. At the same time, we believe this technology can also be developed in ways that deepen human connection and transform relationships in ways that help communities thrive. Our goal will be to foster the latter and mitigate the former.
We believe in a reimagined market capitalism with greater incentives for innovation that expand access to opportunities, resources, and prosperity. We want to encourage investment in areas that can make a difference in people’s lives. Our hope is that generative AI augments human capability and allows us to do more, and better, but not at the cost of wholesale replacement of jobs, industries, or human judgment. Generative AI is a transformative technology that will have a profound impact on the workforce, the economy, and a myriad of industries. Intentional choices today can provide safeguards to secure the future of employment. Again, this is an area that requires immediacy measured with flexibility as we learn more about this emergent technology.
Omidyar Network will specifically reckon with generative AI’s effects on jobs and the future of the workplace to ensure productivity gains are widely shared. We will work to ensure this new technology considers all stakeholders—working people, consumers, and communities. We will help ensure prosperity is distributed more equitably and shareholders are not the only ones that profit from generative AI’s gains.
Digital technology continues to advance more rapidly than laws, regulation, policy, market incentives, and societal norms can keep up. As a result, it has become a continuous struggle to address widespread harms after they have already happened. We, as a society, should learn from the lessons of the past and act before it’s too late with these emergent technologies. Policymakers, regulators, and business leaders have done this before for other complex and multi-purpose technologies, including nuclear and biomedicine. They can do it again.
We believe regulation and innovation can not only co-exist, but when done well, proper governance can incentivize collaboration and mechanisms that lead to world-class advancement. The way forward will require both government and industry to establish well-designed accountability systems with appropriate guardrails and checks and balances to prevent severe harms, establish liability, create public alternatives, and provide remedy and redress.
Omidyar Network will partner with others to determine how to best develop and sustain broad-based governance and checks on power. We are committed to helping address the structural reasons that digital technology escapes meaningful scrutiny. We know from experience this needs to take on several forms, including legislation, regulation, national frameworks, and normative standards. Moreover, we know governance cannot advance without also recognizing the need for global, national, and industrial agreements on norms, codes, and compacts.
Additionally, we will invest in strengthening civil society capacity and expertise to develop, for example, new impact assessment methods and technical audit capabilities. And we will support leaders building out alternatives, such as public options and open-source infrastructure. Omidyar Network will also leverage its for- profit capabilities to invest in efforts that either harness the technology to public interest use or influence how it can have a positive impact on people’s lives.
Generative AI builds upon multiple threads of Omidyar Network’s long-term responsible technology work, including advancing good digital ID, open source ecosystems, platform accountability, privacy and encryption, reimagining data paradigms, new technology models, and trust and safety online.
Generative AI also aligns with Omidyar Network’s initiatives to strengthen worker power, promote stakeholder primacy, and build cultures of belonging, both in terms of who participates in developing technology as well as its profound societal impacts.
Omidyar Network will mobilize our own resources and networks, as well as align with other funders and leaders to develop, share, and realize a nuanced and compelling vision for generative AI.
To take on this incredible task and opportunity, we need so many others to join us. Omidyar Network is committed to working with and coalescing others in philanthropy so we can forge connections with those who can affect change and help realize a vision for generative AI that is grounded in societal good.