Artificial intelligence is transforming every aspect of the online information environment. AI chatbots are becoming increasingly woven into everyday life – from how people search for information, to how they communicate and make decisions, to how they socialize. AI agents are increasingly powering new paradigms of information delivery and software development, calling into question long-established notions about the respective roles of humans and computers in society.
The rapid diffusion of AI to billions of users raises the question of what the future can – or should – look like in an age when machine-generated information stands to become potentially infinite. The early years of the AI revolution have raised serious concerns about safety, authenticity, misappropriation, and concentration of power. But every technological inflection point creates the potential to reimagine what is possible, and after decades of policy and business wrangling, our information environment is ripe for reimagination.
To meet this moment, KGI is infusing policy and technology debates with a wealth of cross-disciplinary research and evidence to address fundamental questions about how to mitigate AI-powered information harms, while simultaneously building an information future that serves the common good. That future – full of independent information sources that humans both consume and produce, backed by sustainable business models, built-in guardrails, and checks on market concentration – can only come about through the evidence-informed development of technical architectures, policy frameworks, market mechanisms, and industry norms.
KGI stands in these cross-currents, convening experts and synthesizing their expertise to inform critical decisionmaking across AI policy and technology debates. Current projects include:
Litigating Technology Design
In recent years, individual plaintiffs and US state governments have brought a growing number of cases against AI companies centered on consumer protection, deceptive practices, and alleged harms associated with chatbot products.
Yet, as several of these cases move into discovery and remedies phases, litigators and courts often lack empirically grounded tools and frameworks needed to address these allegations with effective and durable accountability measures. This creates a gap between rapidly evolving litigation and the infrastructure needed to effectively support discovery, remedy design, and sustained investigation.
KGI’s Litigating Technology Design initiative aims to fill this gap by supporting the implementation of discovery and remedies in social media and AI chatbot cases (see Designing Technology Remedies here), while also simultaneously helping advance the infrastructure necessary for new investigations.
KGI is building on its prior work into the empirical foundations of social media harms to map the emerging evidence base surrounding AI chatbot harms. This work is designed to serve as a resource for litigators, policymakers, and researchers seeking to better understand the real-world impacts of AI systems and to support the development of durable, evidence-based accountability mechanisms. Learn more here.
Agentic Information Future
The development of agentic AI creates the prospect for the internet’s de facto information gatekeepers – social media and search – to be seriously disrupted, if not replaced, by AI agents acting as the new information synthesizers. A world where news and knowledge is increasingly delivered to us proactively by personalized agents, rather than in response to our own searching or scrolling or browsing, raises fundamental questions about every aspect of information production, delivery, and consumption.
Can we sustain the web as we’ve known it, where humans can continue to access news and knowledge directly from original sources, while growing an agentic web of high-quality information in parallel? Can technical and economic mechanisms be built into the agentic web to support better societal outcomes than what the advertising-supported, attention-maximizing web and mobile ecosystems have delivered? Against a backdrop of breathtaking accumulation of wealth in the hands of a small number of giant AI companies, can the design of nascent agentic AI markets build in protections against abuses of market power?
KGI is incubating an Agentic Information Future initiative that will convene experts and deliver insights that are responsive to these questions. The initiative is being designed to clarify this dynamic space and deliver empirical evidence and comprehensive analysis to inform technical and policy debates spanning technology standardization and open source development, competition policy, and media policy.
