Artificial Trailer: Andrew Garfield Steps Into the OpenAI Power Struggle That Ousted Sam Altman

The first trailer for Luca Guadagnino’s Artificial turns OpenAI’s 2023 leadership crisis into a corporate thriller, with Andrew Garfield playing Sam Altman. For Sydney and NSW businesses, its relevance is not celebrity casting. It is the operational risk exposed when board authority, founder influence, staff loyalty, investor dependence and business continuity collide. The episode remains a useful case study in governance design before a leadership dispute becomes an enterprise-wide disruption.

The most important scene in Artificial may not involve artificial intelligence at all.

It may be a boardroom.

The first trailer for Luca Guadagnino’s new film places Andrew Garfield inside one of the most extraordinary corporate leadership crises of the technology era: the November 2023 removal of Sam Altman as chief executive of OpenAI and his return only days later.

Garfield plays Altman. Yura Borisov portrays OpenAI co-founder and former board member Ilya Sutskever, Monica Barbaro plays Mira Murati, and the wider cast includes Ike Barinholtz as Elon Musk and Mark Rylance as Geoffrey Hinton. The film, written by Simon Rich, is scheduled for a limited US theatrical release on 25 December 2026 before expanding in January 2027.

Yet treating Artificial simply as a Silicon Valley biopic would miss what makes the underlying episode commercially significant.

OpenAI's leadership crisis exposed the difference between having formal authority and being operationally capable of using it without destabilising the organisation beneath you.

Editorial note: Artificial is a dramatisation. The analysis below distinguishes the film's presentation from the publicly documented events surrounding OpenAI's November 2023 leadership crisis.

The Trailer Turns Corporate Governance Into The Main Drama

The trailer is deliberately ominous.

Garfield's Altman is presented against the scale of what OpenAI believes artificial intelligence could become. The imagery and dialogue push beyond software development into power, control and responsibility.

That framing is effective cinema, but the documented corporate episode is more precise.

On 17 November 2023, OpenAI announced that Altman would leave as chief executive after its board concluded that he had been "not consistently candid" in communications with the board, which it said hindered the board's ability to perform its responsibilities.

Mira Murati was appointed interim chief executive.

The decision triggered an organisational crisis. Altman subsequently returned, a new initial board was formed and OpenAI began an independent review of what had happened.

The findings published in March 2024 are particularly important because they complicate the simpler story of an ideological battle over artificial intelligence.

OpenAI said the review found a breakdown in trust between the previous board and Altman. It also said the removal was not driven by concerns about product safety, security, development speed, finances or statements to investors, customers or business partners.

More revealingly from an operational perspective, the review found that the decision was implemented on an abridged timeframe, without advance notice to key stakeholders, and that the previous board had not anticipated the extent to which its action would destabilise the company.

That is the business case study hiding underneath the film.

A Board Can Have Authority And Still Create An Operational Shock

Businesses often discuss governance as though authority moves neatly through an organisational chart.

Directors govern. Executives manage. Employees execute. Investors provide capital. Technology partners supply infrastructure. Customers purchase services.

Real organisations are rarely so clean.

Influence can sit outside the formal hierarchy. A founder may hold unusual credibility with employees. A major technology partner may be critical to the operating model. Specialist staff may possess knowledge that cannot be replaced quickly. Customers may depend on uninterrupted services. Investors may have strategic influence even where their formal governance rights are limited.

When those relationships are tightly interconnected, a leadership decision can become an operational event within hours.

CEO or founder removed suddenly

Governance event: CEO or founder removed suddenly
Operational consequence to test: Who immediately receives decision-making authority and customer-facing responsibility?

Board and executives disagree

Governance event: Board and executives disagree
Operational consequence to test: Which decisions can continue and which must pause?

Key staff support one leader

Governance event: Key staff support one leader
Operational consequence to test: How exposed is the business to concentrated knowledge or coordinated departures?

Strategic partner becomes involved

Governance event: Strategic partner becomes involved
Operational consequence to test: Which systems, capital arrangements or services depend on that relationship?

Leadership is reinstated

Governance event: Leadership is reinstated
Operational consequence to test: How are authority, governance controls and stakeholder confidence rebuilt?

This is why the Artificial story should interest operations leaders as much as technology executives.

Governance fails operationally when an organisation knows who can make a decision but has not planned what happens throughout the business once that decision is made.

The OpenAI Crisis Was Also A Dependency Map

Every large organisation has dependencies.

The question is whether management can see them before a crisis activates several at once.

For an artificial intelligence company, those dependencies can include specialist researchers, cloud infrastructure, capital, model deployment systems, intellectual property, regulatory relationships, corporate reputation and access to highly competitive technical talent.

For a Sydney property, construction or professional-services operator, the names are different but the structure is familiar.

A project may depend simultaneously on:

a project manager who holds the client relationship;
a specialist contractor whose availability controls programme timing;
a strata approval or building-access window;
a supplier with a long material lead time;
a software platform holding project records;
a finance approval needed before procurement;
a principal contractor or consultant controlling site sequencing; and
a small number of people who understand how all of those pieces connect.

Remove one apparently central person or approval point without mapping the downstream consequences and the business can discover that its formal structure was never its real operating structure.

Founder Risk Is Not The Same As Key-Person Risk

Artificial also arrives at a time when many technology businesses remain closely identified with individual founders and chief executives.

That creates two different risks.

Key-person risk is the practical difficulty of losing someone whose skills, knowledge, customer relationships or authority are difficult to replace.

Founder-dependence can be broader. The person may also represent the company's strategy, employee confidence, fundraising narrative, customer trust and external identity.

A board may therefore have the legal or constitutional authority to change leadership while discovering that large parts of the organisation have developed around a person rather than a role.

The solution is not to make founders untouchable.

It is to build an operating model capable of surviving disagreement at the top.

What Sydney Businesses Should Document Before A Leadership Crisis

Australian company directors are not ceremonial supervisors. ASIC's guidance for company officeholders makes clear that directors are responsible for understanding and overseeing the business.

That responsibility becomes more difficult where strategic authority and operational knowledge are concentrated in different places.

A practical continuity review should therefore go beyond the succession-plan document stored with corporate records.

Map decision authority. Identify who can approve payments, contracts, hiring, customer commitments, system changes, public communications and project variations.
Map operating dependencies. Record which people, platforms, contractors, suppliers and external partners are essential to continuing service.
Identify concentrated knowledge. Determine where only one or two people understand a major system, commercial relationship or operational process.
Define temporary authority. Decide who takes control if a chief executive, director, project leader or system owner becomes unavailable.
Create stakeholder sequencing. Determine which employees, partners, customers, advisers and regulators require communication before or immediately after a material leadership event.
Test continuity rather than assuming it. Run scenarios in which a key executive, system or supplier becomes unavailable with little warning.

The objective is not predicting a dramatic boardroom confrontation.

It is making sure that an unexpected decision at the top does not force the rest of the organisation to improvise its authority, communications and service continuity in real time.

The AI Governance Lesson Is Different From The Usual AI Governance Debate

Much of the current discussion around enterprise artificial intelligence focuses on data access, privacy, model accuracy, agent permissions and human approval.

Those remain important. Elyment has examined why AI governance needs to move ahead of production deployment, how real enterprise AI adoption spreads through organisations, and what changes when AI implementation becomes a large-scale consulting and operational discipline.

Artificial raises a different layer of governance.

What governs the organisation that governs the AI?

A technically well-controlled AI platform can still sit inside a company with unclear executive authority, concentrated knowledge, fragile stakeholder relationships or inadequate business-continuity planning.

AI governance therefore cannot end at the model.

It also has to cover the organisation responsible for deciding what the model is allowed to do.

NSW Is Moving Towards More Explicit AI Accountability

NSW provides a useful example of how institutional governance around AI is becoming more structured.

The NSW AI Assessment Framework requires NSW Government agencies to identify AI risks, define mitigations and establish governance throughout the system lifecycle. High and critical-risk uses are subject to additional review.

The framework is designed for government and should not be presented as a private-sector compliance requirement.

Its underlying operating principle is nevertheless relevant to Sydney businesses: accountability should be identifiable before a high-consequence system or decision reaches production.

The same principle can be applied to corporate governance.

If a board can remove a chief executive, who owns operational continuity afterwards?

If an executive can authorise an AI system, who can suspend it?

If a technology partner becomes business-critical, who manages that dependency?

If staff reject a strategic decision, who controls client delivery while the dispute is resolved?

Governance becomes useful when those answers exist before the incident.

A Better Resilience Test For Founder-Led And Technology-Dependent Businesses

Businesses do not need an OpenAI-sized controversy to experience the same structural problem.

A six-part test can reveal much of the exposure.

Authority

Control area: Authority
Question for management: Can the business identify who has power to make each material decision today?

Succession

Control area: Succession
Question for management: Can another person assume the role without losing access, information or commercial authority?

Dependencies

Control area: Dependencies
Question for management: Which people, suppliers, platforms or investors could materially disrupt operations?

Communication

Control area: Communication
Question for management: Who must be informed, in what sequence, if leadership changes unexpectedly?

Continuity

Control area: Continuity
Question for management: Which customer, project and system activities must continue regardless of leadership uncertainty?

Recovery

Control area: Recovery
Question for management: How are trust, authority and operating controls restored after the immediate crisis ends?

A company that cannot answer those questions may have an organisational-chart view of governance rather than an operational one.

Why The Story Matters More In 2026 Than It Did In 2023

When OpenAI's crisis occurred in 2023, generative AI was still moving from public experimentation into widespread enterprise adoption.

By 2026, artificial intelligence increasingly sits inside document systems, coding environments, customer workflows, cybersecurity operations, financial processes and enterprise software.

The organisations building and operating those systems therefore have consequences far beyond their own offices.

Leadership stability, governance design and business continuity at major AI providers can affect customers, developers, cloud partners, investors and businesses whose own operations depend on their services.

That makes the governance of an AI company part of the broader AI-risk conversation.

The technology cannot be separated entirely from the institution controlling it.

The Most Useful Question Artificial Leaves Businesses With

Audiences will ultimately decide whether Guadagnino's film succeeds as drama.

Businesses should ask a different question.

What would happen to our organisation if an equivalent leadership shock occurred tomorrow?

Would everyone know who was authorised to act?

Would critical projects continue?

Would system access remain controlled?

Would customers receive consistent information?

Would strategic partners know who represented the company?

Could the board make a major decision without accidentally discovering that the organisation's real power structure sits somewhere else?

The OpenAI episode demonstrated that governance is not only about whether a decision is permitted.

It is also about whether the organisation is prepared to absorb the consequences of that decision.

GOVERNANCE · OPERATIONS · BUSINESS CONTINUITY

Review The Operating Model Before A Critical Decision Tests It

Map decision authority, key-person dependencies, AI governance, stakeholder handovers, system ownership and business-continuity controls before an executive, technology or project disruption becomes a wider operational problem.

Request An Operational Project Review

The Bottom Line

Artificial turns the 2023 OpenAI leadership crisis into cinema, with Andrew Garfield taking on one of the technology industry's most scrutinised executives.

The public record is more useful to businesses than the mythology.

OpenAI's former board had authority to change leadership. What followed showed that authority alone does not guarantee organisational control. Leadership, employees, strategic partners, governance arrangements and business continuity were deeply interconnected.

For Sydney and NSW organisations, that is the practical lesson.

Strong governance should not merely determine who can make a decision. It should define how the business continues operating once that decision creates consequences across people, systems, customers, projects and partners.

Sources And References

GOVERNANCE · OPERATIONS · BUSINESS CONTINUITY

Review The Operating Model Before A Critical Decision Tests It

Request a Project Review