In the space of just three days this month, two very different AI companies delivered the same uncomfortable lesson about leadership, incentives, and accountability.
On July 10, Apple filed a lawsuit in the U.S. District Court for the Northern District of California against OpenAI, its hardware subsidiary io Products, Chief Hardware Officer Tang Tan, and former Apple engineer Chang Liu. The complaint alleges a coordinated campaign to misappropriate Apple’s trade secrets to accelerate OpenAI’s consumer hardware development.
Tang Tan is a 24-year Apple veteran who rose to become Vice President of Product Design, where he oversaw hardware engineering for the iPhone and Apple Watch. After leaving Apple in early 2024, he co-founded io Products alongside legendary designer Jony Ive to develop AI-native consumer devices. OpenAI acquired io Products in a major deal, and Tan was named Chief Hardware Officer, leading the company’s hardware efforts and reporting directly to CEO Sam Altman.
According to Apple’s complaint, Tan allegedly used his knowledge of Apple’s internal projects and security processes to direct aggressive recruiting tactics. The lawsuit claims he instructed recruiters and interviewers to ask Apple employees still at the company to bring unreleased hardware components to interviews for “show and tell” sessions. The complaint further alleges that Tan coached departing Apple employees on how to evade Apple’s security procedures when leaving the company. It also accuses another former Apple engineer, Chang Liu, of failing to return his company laptop after joining OpenAI and allegedly using it to access Apple’s internal network from OpenAI’s offices, downloading sensitive technical documents.
Apple claims that OpenAI approached one of its exclusive manufacturing partners and requested they perform a proprietary metal-finishing process, while falsely claiming they had Apple’s permission. The complaint states that Apple had previously sent a letter to OpenAI in February 2026 raising concerns about confidential information potentially making its way to OpenAI, but received no response. Apple describes OpenAI’s hardware business as resting on a foundation “rotten to its core” due to this alleged reliance on misappropriated trade secrets. The complaint notes that more than 400 former Apple employees now work at OpenAI.
It is also worth noting that Tang Tan had already been away from Apple for more than two years when these allegations emerged. This raises an important question: to what extent might the culture and environment he entered at OpenAI — a fast-moving, high-stakes organization aggressively building hardware capabilities — have influenced the alleged behavior? Even experienced and previously successful leaders can be shaped by the incentives, norms, and pressures of the organizations they join. Company culture is not neutral. It either actively reinforces ethical behavior or creates conditions where cutting corners, normalizing questionable practices, or prioritizing speed above all else becomes more likely.
This dynamic — where organizational culture and competitive pressure can influence even seasoned leaders — is not unique to one company. Just three days earlier, on July 7, Meta launched Muse Image, its most advanced image generation model developed by Meta Superintelligence Labs under Chief AI Officer Alexandr Wang. The model was positioned as an agentic system capable of following complex instructions, editing images with precision, composing from multiple references, and drawing on Instagram for social context. As part of the launch, Meta introduced a feature that allowed users of Meta AI to generate images by @-mentioning public Instagram accounts, pulling photos and likenesses from those profiles
Caption: Choosing the ethical path in AI leadership requires cultures that can balance internal competitive pressure with external trust and long-term societal legitimacy.
For public accounts, this capability was enabled by default. Users did not need to give consent, and they received no notification when their content was used. Only private accounts and accounts belonging to users under 18 were automatically excluded.
This design was particularly flawed because many users deliberately choose public accounts to gain visibility, build an audience, network, or advance their careers — not because they are comfortable with anyone using their photos and likeness to generate AI images. By making the feature opt-out by default for public accounts, Meta effectively created a false choice: stay visible and risk non-consensual use of your likeness, or go private and lose reach and opportunity. The feature quickly drew sharp criticism from privacy advocates, SAG-AFTRA (the Screen Actors Guild – American Federation of Television and Radio Artists), the Creative Artists Agency (CAA), and members of the public who raised concerns about non-consensual use of likeness, potential harassment, deepfakes, and the exploitation of creative work. On July 10, Meta removed the specific @-mention feature, stating that it had “missed the mark” and that its intent had been to provide a “useful creative tool.”
Two different failures, one shared warning
While both incidents are serious, they represent different kinds of leadership and organizational failure.
The allegations in the Apple lawsuit, if true, go beyond errors in judgment. They describe a pattern of deliberate actions — allegedly directed from the level of a Chief Hardware Officer — to extract confidential information and trade secrets from a competitor. If proven, this would not be a case of moving too fast or misreading the environment. It would represent a conscious decision to cross ethical and legal boundaries in pursuit of competitive advantage. That is a fundamentally different category of failure: one that, if substantiated, reflects unacceptable behavior at a leadership level rather than a miscalculation.
Meta’s Muse Image episode, while damaging, sits in a different category. It appears to be primarily a failure of judgment about the environment. Meta was clearly trying to close the gap in generative AI capabilities and gain competitive positioning in image and creative tools. In doing so, the company appears to have significantly misread the sentiment of its user base, creators, and the broader public regarding consent, likeness rights, and the acceptable use of people’s personal content in AI systems. The decision to make the feature opt-out rather than opt-in, combined with the lack of notification, suggests leadership underestimated how strongly users and advocacy groups would react to what felt like an extractive default. This was a serious misjudgment of the cultural, social, and political environment around AI in mid-2026 — but it was still fundamentally a judgment error, not an alleged intentional transgression of ethical or legal lines.
Both cases are nevertheless deeply concerning, and both point to the same underlying problem: when the financial stakes are measured in billions, the pressure to move quickly and aggressively can distort decision-making at multiple levels.
Why these incidents matter beyond Silicon Valley
There is a larger reason these episodes constitute a wakeup call — not just for the companies involved, but for Silicon Valley and the broader model of democratic capitalism it represents.
Democratic capitalism has delivered unprecedented innovation, prosperity, and individual freedom precisely because it combines market incentives with democratic accountability and (ideally) ethical guardrails. Its long-term legitimacy, however, does not rest solely on economic performance. It also depends on the public’s perception that the system produces broadly fair outcomes and that its most powerful actors — especially in technology — generally operate within ethical and legal boundaries.
When high-profile technology companies appear to prioritize speed, competitive positioning, or growth at the expense of consent, fairness, or legal norms, it weakens that perception. Over time, this erosion of trust makes alternative models — whether heavier state control, more socialist-leaning approaches, or highly centralized regulatory regimes — more appealing to segments of the public and policymakers. In other words, the continued success and societal acceptance of democratic capitalism is partly contingent on whether its leading institutions are seen as capable of self-regulation and ethical leadership.
Silicon Valley currently sits at the center of this dynamic. As the most visible engine of innovation-driven capitalism in the world, the behavior of its major companies has an outsized influence on how millions of people — and governments around the globe — judge the fairness and legitimacy of the system itself. When trust erodes, the political and regulatory environment becomes more hostile, talent becomes harder to attract and retain, and the freedom to operate that has fueled decades of progress narrows.
This is why incidents like these are not merely corporate missteps. They are tests of whether the ecosystem that has driven so much of the Fourth Industrial Revolution can maintain the ethical credibility required to sustain broad societal support for the model that makes it possible.
Why distributed values-based leadership is now essential
This is why the core argument in my book AI L4IR: Leadership and the 4th Industrial Revolution has become increasingly urgent. In this era, ethical leadership cannot be concentrated at the C-suite level or delegated primarily to legal and compliance teams. It must be highly distributed — embedded in the daily decisions and culture of the entire organization.
When the financial stakes are this high, it is unrealistic to expect a small group at the top to identify and mitigate every ethical risk. Real protection comes when engineers, product managers, recruiters, and middle managers are both expected and supported to ask “Should we do this?” — and when the organization has mechanisms that allow those questions to be heard and acted upon without fear of slowing down progress.
Distributed values-based leadership requires more than statements of principle. It requires building systems in which:
Ethical considerations are treated as core product and technical requirements from the earliest stages of development.
People at all levels are recognized and rewarded for raising difficult questions, not just for shipping quickly.
Recruiting, talent, and partnership practices have clear, actively enforced boundaries — even when competitive pressure is intense.
Leadership at every tier consistently models that long-term trust and integrity are non-negotiable, even when short-term gains appear attractive.
The organizations that treat values-based leadership as a distributed operating system — rather than a slogan or a top-down policy — will be far better positioned to navigate the extraordinary financial and competitive pressures of this era while preserving the trust required to lead. Those that do not will continue to discover the costs of getting it wrong — not only for themselves, but potentially for the broader system that has enabled their success.
Endnotes
Apple Inc. v. OpenAI Foundation et al., U.S. District Court for the Northern District of California, Case No. 5:26-cv-07078, Complaint filed July 10, 2026. The complaint alleges trade secret misappropriation and breach of contract involving former Apple employees, including Chief Hardware Officer Tang Tan.
Meta Platforms, “Introducing Muse Image and Muse Video,” official blog post, July 7, 2026. Meta announced its new agentic image generation model developed by Meta Superintelligence Labs.
Meta Platforms statement on removal of the @-mention feature, July 10, 2026: “We’ve heard the feedback that this feature missed the mark, so it’s no longer available.”
Reporting on the Muse Image controversy and backlash:
TechCrunch, “Meta just launched a new AI generator, and users are already pushing back,” July 7, 2026.
WIRED, “Meta Now Lets Anyone Use Your Instagram Photos in AI Images—Unless You Opt Out,” July 7, 2026.
The Hollywood Reporter and Axios coverage of SAG-AFTRA and CAA statements (July 8–10, 2026).
Details on the public vs. private account default settings and user backlash are drawn from contemporaneous reporting in TechCrunch, WIRED, and Meta’s own help documentation (July 2026).
Background on Tang Tan’s career at Apple drawn from Bloomberg reporting (December 2023) and subsequent coverage of his departure and role at OpenAI.