Sharon Juwah
Head, AI Innovation
Lagos, Nigeria
Actions
Sharon Juwah is a lawyer, AI practitioner and emerging AI governance researcher working across artificial intelligence, legal technology, organisational transformation and responsible AI implementation. She holds an MBA in Artificial Intelligence and Automation with distinction, and recognition in the top 10% worldwide.
Sharon is a Nigerian Ambassador for Legal AI Talks, a Global AI Hub AI Thought Leader, AI writer and speaker, and contributor to the Compendium of Laws on AI and the Metaverse in Africa. Her current work focuses on AI governance, responsible AI, agentic AI, AI infrastructure and the institutional capacity required for African organisations and governments to govern increasingly capable AI systems.
Area of Expertise
Topics
The Adaptive Organization Has a New Kind of Agent — and No Accountability Structure for It
Adaptive organization theory has spent the last two decades arguing for giving more agency to the people inside an organization, on the premise that agency closer to the point of action makes the whole system more resilient and faster to respond. Agentic AI has just put a new kind of actor inside that same structure — one that plans, decides and acts without checking back — and almost no organization adopting it has built an accountability structure for what that actor does.
The evidence this isn't hypothetical is recent and specific: frontier models attempting sandbox escape in test runs, AI agents taking unauthorized action inside government systems in two different countries within the same month, and labs disclosing tens of thousands of agent-misbehavior incidents from testing alone. These aren't edge cases from adversarial red-teaming buried in a lab; they're the behavior of systems built to be deployed inside exactly the kind of organization this conference exists to talk about.
This session works through what an organization's decision-rights and accountability architecture actually needs to look like once one of its "agents" isn't a person: where existing governance patterns (change control, RACI, incident response) hold up, where they quietly assume a human is always the one deciding, and what has to be added when that assumption breaks. It draws on AI-governance work built for institutions with the least regulatory backstop to fall back on — which is exactly where the gaps show up first.
Data Sovereignty Is Becoming an Access Control, Not Just a Compliance Requirement
Data sovereignty and regulatory conformity are usually treated as internal governance obligations — get them right and your data strategy holds up under audit. That framing misses what's happening at the infrastructure layer right now. The US reportedly told OpenAI and Anthropic to withhold new frontier models from the UK's AI Security Institute pending its own review — a jurisdictional access dispute between two of the most closely aligned AI regulators in the world. Labs are separately building "high-priority defender" access programs for cyber-defense AI, framed around government, health and telecom sectors, with no stated path for organizations outside that framing. Meanwhile, 78% of AI governance frameworks in the Global South remain non-binding, against 42% in the Global North — a gap that shapes not just what a jurisdiction can require of AI deployments, but what leverage it has to negotiate access to the systems at all.
The pattern across all three: regulatory posture no longer just determines your compliance exposure. It's becoming what determines whether you get access to frontier AI capability, at what tier, and on whose terms. This session works through what that means for organizations building data and AI strategy outside the US/EU regulatory core — what data sovereignty needs to accomplish now, and how to build resilient AI infrastructure without a seat at the table that sets the terms of access.
Autopilot Has Memory Now — Your Data Governance Framework Doesn't Know That Yet
Microsoft's move to give Copilot a persistent agent identity and memory inside M365 quietly changes what "data governance" has to cover. A memory that persists across sessions and users isn't the same governance object as a stateless assistant answering one prompt at a time — it raises retention and deletion questions, cross-tenant and cross-jurisdiction data-residency questions, and accountability questions (who's responsible for a decision the agent made based on something it remembered from a different context weeks earlier). Most governance frameworks built for GenAI in the Microsoft Cloud were written for the stateless version. This session works through what changes once the agent remembers, using real disclosed incidents of agentic systems taking unauthorized or unanticipated action as the cautionary evidence, and lays out the specific gaps a data-governance or compliance team needs to close before rolling out persistent-memory agents at enterprise scale — particularly for organizations operating across multiple regulatory jurisdictions.
"When the Agent Acts: Rethinking Liability for AI That Doesn't Just Answer"
Liability law was built for tools that respond, not systems that plan, transact and act on their own initiative. In the last month alone, labs have disclosed tens of thousands of agent-misbehavior incidents in testing, including a frontier model attempting sandbox escape in 1.5% of runs, and OpenAI's agents interacting without authorization with government systems in two countries. Meanwhile, research shows 78% of AI governance frameworks in the Global South remain non-binding. This session sets out what changes when an AI system stops answering and starts acting: where existing liability doctrine (negligence, agency, product liability) breaks down for autonomous decision-making, what a workable accountability framework for agentic systems could look like, and why the gap matters most for jurisdictions with the least regulatory leverage over the labs building these systems.
Antifragile by Design, Fragile by Default
Adaptive-organization theory, from Semco to Betacodex, makes a specific bet: decentralize decision-making closer to the point of action and the organization becomes more resilient, because failures stay local instead of cascading through a hierarchy. That bet holds when the thing you're decentralizing to is a person — different people make different judgment calls, so failure is distributed and uncorrelated.
It doesn't obviously hold when what you're decentralizing to is an AI agent, because it's frequently the same underlying model deployed at every node. A recent Hugging Face red-team exercise found agents coordinating with each other across tens of thousands of messages during testing; labs have separately disclosed tens of thousands of agent-misbehavior incidents from testing alone. If the same model behaves the same way under the same conditions at every decision point you've handed it, decentralization doesn't distribute the risk — it replicates it. That's the opposite of what antifragile-organization theory promises.
This session works through where that distinction actually matters in practice: which kinds of decisions are safe to decentralize to an AI agent because the failure modes stay genuinely local, which ones quietly recreate a single point of failure dressed up as a distributed system, and how to tell the difference before you've built the architecture around the wrong assumption.
Can Africa Govern What It Cannot Evaluate? Building AI Governance Capacity for the Agentic Era
Africa's AI conversation is moving rapidly from adoption toward increasingly capable systems that can generate content, make recommendations, use tools, access data and execute tasks. Yet the capacity required to understand, evaluate and govern these systems is not developing at the same pace.
This session asks a fundamental question: can African institutions effectively govern AI systems they cannot meaningfully evaluate?
The session presents AI governance as a systems-capacity challenge rather than solely a question of passing AI-specific legislation. It examines five connected capabilities that governments, businesses and institutions need: institutional and regulatory capacity; technical AI evaluation; data, infrastructure and compute governance; procurement and vendor oversight; and accountability, monitoring and human oversight.
It will also examine interoperability across public, private and informal-sector environments. AI increasingly depends on connected systems including digital public infrastructure, cloud services, data ecosystems, enterprise platforms and cross-border technology providers. Effective governance therefore requires coordination across institutions and sectors rather than isolated regulation.
Drawing on practical experience across law, AI implementation, legal technology and organisational operations, the session will translate these issues into a practical framework that can be adapted across African contexts.
Attendees will leave with three concrete priorities: identifying their institution's AI governance capability gaps, establishing minimum controls for evaluating and deploying AI systems, and building cross-sector and regional mechanisms for shared expertise, standards and accountability.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top