Addressing AI, Quantum, Human Agency And P(doom) Risk - Forbes

Boards and c-suites increasingly recognize that artificial intelligence and quantum computing can no longer be treated as future issues. Both technologies are reshaping risk, resilience and competitive advantage today. At the same time, debates about “p(doom),” the probability that advanced technologies could create catastrophic outcomes, have moved into mainstream boardroom conversations.

For most organizations, the key question is how best to understand and inventory the ways these technologies can affect critical decisions, business models, talent requirements and risks. Successful organizations are focusing more on governance, resilience and human agency, than on abstract probabilities.

What is p(doom)?

WTW’s Hélène Galy, Jen Daffron and Lucy Stanbrough explain in Beyond p(doom): How Human Agency is Shaping Real-World AI Outcomes that the term is shorthand in contemporary technology discussions for the probability that advanced AI could ultimately cause catastrophic or existential harm. As AI advances and timelines to artificial general intelligence remain debated, p(doom) has become a common reference point across technical, academic, business and policy communities. While the concept can help frame the discussion, the report encourages boards and senior leaders to develop a deeper understanding of how technology-driven risks emerge, spread and affect organizational objectives.

AI is transforming risk

As this space has covered, concern over technology risk has shifted from theory to practice as AI adoption accelerates. Questions of accountability, liability, insurability and governance are critical business issues. Regulatory frameworks are evolving, litigation is emerging and organizations are discovering that AI-related losses exist across multiple areas of responsibility.

Recent analysis by WTW’s Sonal Madhok and Dr. Anat Lior, including their work on Insuring the AI Age, highlights how AI is reshaping risk profiles across industries. Many organizations rely on a patchwork of policies, with no single policy covering the full spectrum of potential losses. Depending on the circumstances, exposures may involve cyber, professional liability, directors’ and officers’ liability, employment practices liability, general liability or other lines of cover.

This complexity creates a governance challenge. Boards and c-suites need to understand where coverage gaps exist, who is accountable and how AI decisions could create unintended liabilities. The largest exposures often come from routine decisions made without sufficient oversight, documentation or controls.

AI risk is not one thing

AI remains difficult to govern because organizations often discuss it as a single thing (and risk category). However, it actually creates a range of risks and challenges, including unreliable outputs, human error and malicious use, technical failures that can extend across interacting systems, broader social and economic disruption, and situations where governance and oversight fail to keep pace with technological change.

For many boards and senior leaders, “Which technology-related failure modes already interact with our business model, and how could they escalate under stress?” is a more relevant question than, “What is our p(doom)?” Effective governance focuses on accountability, ownership, resilience and decision-making.

Quantum computing presents a similar governance challenge

As WTW’s Mark Beardall observes in U.K. National Preparedness Commission, quantum computing has moved beyond the research laboratory and closer to the mainstream. Its potential to transform industries and solve problems beyond today's computing power also creates serious new risks, including undermining the encryption that keeps online communications and data safe.

The most widely discussed concern is “Q-Day” or “Y2Q”: the point at which sufficiently powerful quantum computers can break widely used encryption standards. Even if timelines remain uncertain, information stolen today could be stored and decrypted later when quantum capabilities mature, the “harvest now, decrypt later” scenario.

The leadership question goes beyond simply when Q-Day will occur to include what information must remain secure for the next decade or longer, whether current cyber strategies are sufficient and how to mitigate risks.

The real challenge is interconnected risk

Future disruptions are increasingly likely to emerge through interactions between technologies and existing vulnerabilities. AI-enabled cyberattacks, quantum threats to encryption, geopolitical instability, third-party dependencies, regulatory shifts and critical infrastructure exposures combine in ways that amplify impacts beyond any single technology risk.

However, technology risks often are assessed independently, while real-world failures propagate across interconnected systems. Leaders benefit from understanding how technology-driven risks interact and impact the wider business ecosystem.

Human agency remains the critical differentiator

Technology alone does not determine outcome. Outcomes are shaped by the choices human leaders make about governance, controls, accountability and adoption, and by people executing those decisions.

The organizations most resilient to emerging technology risks understand, challenge, intervene and adapt when tools fail. Research by the Willis Research Network points to four areas where leaders can strengthen resilience:

- Technical controls: To what extent can the organization meaningfully constrain, monitor and override AI behavior in live environments? Strong controls help organizations identify and manage hallucinations, drift, misuse and model failures before they become costly. In terms of quantum computing, this means creating a plan and understanding where encryption is being used and how it operates.

- Regulation and governance: AI raises questions around accountability, transparency, fairness, documentation and customer impact. Organizations that move beyond compliance to create trustworthy AI can gain a competitive advantage.

- Knowledge and capability: Capability gaps create control weaknesses. Organizations strengthen resilience when professionals maintain the expertise necessary to challenge AI outputs and recognize errors.

- Adoption strategy: Technology that is widely used but poorly understood creates unmanaged exposure. Leaders benefit from knowing which critical processes depend on AI and what contingencies exist if those systems fail.

Questions boards should ask

Amid those challenges, boards would benefit from asking:

- Which strategic assumptions would fail under plausible AI or quantum-related scenarios?

- Which critical business processes would fail if these technologies became unavailable, compromised or unreliable?

- Which combinations of technology-related risks could push us beyond our company’s risk tolerances?

- Who is accountable for converting technology insight into decisions and action?

From predicting the future to shaping it

Discussions about p(doom), existential AI risks and quantum disruption can create the impression that technological risk will arrive through a single dramatic event. However, the most significant impacts are more likely to emerge through everyday decisions: a model trusted too readily, a dependency overlooked, a control weakened or a technology adopted before it could be sufficiently governed.

The future may be uncertain, but effective leaders use uncertainty as a catalyst for action. They identify plausible scenarios, understand their implications and build the capabilities needed to respond. Their objective is to ensure their organizations remain resilient, adaptable and prepared to act under uncertainty.