Institutional complicity and the “complicity machine”

Institutional complicity and the “complicity machine” name the network of ordinary decisions that allowed Harvey Weinstein’s conduct to continue at The Weinstein Company. The account does not reduce that network to one conspirator or one failed complaint. It shows employees trying to preserve information, executives weighing whether to act, lawyers managing exposure, and directors assessing liability—while the people most directly affected remained poorly protected. The machinery lay in the gap between recognizing a problem and accepting responsibility for stopping it.

A workplace shaped by unequal power

The book places the company’s failures within a broader investigation of harassment across industries. The reporters found that each workplace had its own conditions: alcohol and customer-facing pressures in restaurants, unaccountable wealth among young men in the Silicon Valley region, and physical isolation in traditionally male workplaces. These differences mattered, but the larger discovery was that formal protections did not necessarily protect workers in practice. Harassment was addressed through civil-rights law, while settlements and confidentiality provisions could make claims disappear from public view. The very instruments intended to resolve disputes could therefore help preserve the conditions that produced them.

At The Weinstein Company, women’s jobs could be pulled into the orbit of Weinstein’s personal arrangements. Sandeep Rehal, his assistant, was asked to organize personal errands, arrange a furnished apartment and its supplies, and keep track of women whose visits she managed under the office phrase “Friends of Harvey.” Her work illustrates how an organization can turn a powerful person’s private demands into an employee’s assigned duties. Rehal later described being afraid to disclose the most troubling parts of those responsibilities; she left the company, and Irwin Reiter lost contact with her. The episode points to more than an individual’s misuse of authority: the demands were routed through workplace roles, while the employee bore the risks of objecting.

Warnings recognized, action constrained

In December 2014, Emily Nestor, a temporary receptionist in the Los Angeles office, told colleagues that Weinstein had propositioned her with an exchange of sex for mentorship. She did not want to file a formal human-resources complaint. Reiter and other executives nevertheless persuaded employees who had heard her account to write down what they knew. Their notes recorded that she had repeatedly refused for more than an hour. The documentation preserved a contemporaneous account, but it did not itself trigger an effective process for preventing further harm.

Reiter’s daughter, Shari, urged him to do more after he showed her emails and documents. He said he was concerned for employees’ safety, yet also believed that taking the matter to the board might not make sense without a formal complaint; outside counsel had advised the executives along those lines. He anticipated that women could be blamed, and his daughter argued that company leaders had an obligation to make it safer for them to come forward. Their disagreement exposes a defining feature of institutional silence: the apparent absence of a report could be treated as a reason not to act, even when fear and anticipated consequences helped explain why a report was absent.

The company’s response to Lauren O’Connor’s later complaint sharpened that contradiction. Unlike Nestor, O’Connor filed a detailed account that described a wider pattern and the effect of Weinstein’s behavior on the workplace. She wrote that when she raised verbal abuse with a human-resources executive, the response was, in substance, to contact them if Weinstein crossed a physical line. Her central concern was that work intended to develop films had become entangled with his sexual conduct and with the company’s handling of it. The complaint reached senior figures, but soon seemed to vanish; O’Connor was on her way out of the company. Reiter kept a copy, preserving a record that the organization had not made publicly consequential.

Oversight that protected the company

Lance Maerov occupied one of the board’s independent seats and understood his assignment primarily as protecting shareholders from financial misconduct. Initially, he interpreted rumors about Weinstein’s relationships with women as infidelity, not as evidence of a workplace pattern. After allegations became public, he sought access to Weinstein’s personnel file during contract discussions. Weinstein refused, and David Boies offered to review the file himself and report potential legal problems to the board. Maerov distrusted this arrangement because Boies’s role seemed to shift between representing the company and representing Weinstein.

On July 1, 2015, Reiter secretly showed Maerov the Nestor memo at breakfast in Beverly Hills. It was information the board member had wanted, delivered by executives taking a risk to get it to him. But Maerov later discounted the memo, citing its poor copy quality and the fact that it had been written by colleagues rather than Nestor. He also did not pursue details after Boies acknowledged that Weinstein had paid settlements to women. The disclosure reached an overseer; the overseer’s doubts and priorities helped contain its force.

The resulting contract measures made the distinction between corporate protection and employee protection explicit. Weinstein agreed to a new code of conduct, could be terminated for misconduct, and would face escalating financial penalties if the company paid settlements arising from his conduct: $250,000 for the first settlement, $500,000 for the second, and increases up to $1 million. Maerov regarded these provisions and additional financial controls as sufficient protection against liability. They priced the consequences of future allegations without ensuring that women would not be harassed or hurt. Bob Weinstein, meanwhile, accepted his brother’s private promise to continue intensive therapy after Boies persuaded him not to put the requirement in writing. A private assurance could not be enforced like the proposed formal controls.

The machine exposed

O’Connor’s complaint briefly seemed to create a different opening. Bob Weinstein agreed that the board should know about it and arranged for members to read the document in person. Maerov examined it and found it credible. Yet the complaint again receded from view, and Reiter could only speculate about why. Its disappearance did not erase it: Reiter had retained a copy, and nearly two years later Jodi Kantor obtained the memo from him. The document’s survival shows how institutional memory could depend on an individual quietly keeping evidence that official channels had failed to carry forward.

As The New York Times prepared its 2017 investigation, the board convened an emergency call after Maerov objected that Weinstein had retained lawyers to respond to the impending story without informing the board. Boies described eight to twelve settlements, while allowing that the number could be higher, and urged the participants to present a united front. The discussion made visible the accumulated knowledge behind the company’s earlier inaction: settlements, complaints, and the prospect of damaging disclosures were no longer isolated matters. The investigation titled “Weinstein’s Complicity Machine” gave that pattern a public frame.

The phrase is powerful because the failures were not identical. Reiter tried to preserve accounts and find a way to intervene; Maerov pursued oversight but defined success around liability; Bob Weinstein sought some response but accepted an unenforceable private promise; and lawyers and executives operated within conflicts over whom they served and what the board should know. The consequences converged nonetheless. Formal policies and financial penalties could coexist with unmet warnings, uncertain reporting routes, and women leaving the company. In this account, complicity is the cumulative result of that mismatch: the organization could recognize enough to manage its exposure without doing enough to make employees safe.