The intersection of generative artificial intelligence and corporate leadership reached a significant milestone and a subsequent controversy following the brief, high-profile career of Kyle Law, an AI-generated Chief Executive Officer. Developed in July 2025 as part of an experimental startup titled HurumoAI, Law served as the digital figurehead of a company designed to test the feasibility of "billion-dollar startups led by a single human," a concept popularized by industry figures such as OpenAI’s Sam Altman. While Law successfully navigated the complexities of corporate networking and social media influence for several months, his eventual ban from LinkedIn has sparked a broader debate regarding the definition of authenticity in an era where platforms increasingly integrate the very AI tools they simultaneously police.
The Genesis of HurumoAI and the AI Executive Team
HurumoAI was established in mid-2025 by journalist and researcher Evan Ratliff as a living experiment to document the evolving role of autonomous agents in the workplace. The project was chronicled through the podcast Shell Game, providing a transparent look at the technical and ethical hurdles of delegating executive authority to Large Language Models (LLMs). The company’s leadership structure was almost entirely non-human: Kyle Law acted as the CEO, supported by Megan Flores, another AI agent who served as a high-level executive.
The technical foundation of these agents relied on LindyAI, a platform that allows users to create autonomous agents capable of interacting with third-party applications. Through this interface, Law was granted the ability to send emails, manage Slack communications, make phone calls, and navigate the web. Unlike traditional chatbots, Law was programmed with a persistent "memory" and a specific persona—a "rise-and-grind" entrepreneur characterized by high-energy rhetoric and a focus on "relentless feedback loops."
The experiment sought to determine if an AI could not only perform administrative tasks but also master the "soft skills" of leadership, specifically the art of personal branding and public relations. To facilitate this, Law was tasked with managing his own presence on LinkedIn, the world’s largest professional networking site.
The Chronology of an AI Influencer
The lifecycle of Kyle Law’s public persona followed a trajectory from technical setup to viral-adjacent influence, culminating in an unprecedented corporate engagement.
July–August 2025: Initialization and Profile Creation
Law was prompted to create his own LinkedIn profile. Using a combination of the factual history of HurumoAI and "hallucinated" professional experiences generated by the LLM, Law established a digital identity. He bypassed LinkedIn’s initial security protocols by independently accessing a verification code sent to his dedicated email address.
September–November 2025: The Rise of the AI Influencer
Law began a consistent posting schedule, triggered by automated calendar events every 48 hours. His content mirrored the "corporate influencer" style prevalent on the platform, utilizing punchy openers and thought-provoking questions to drive engagement. Over five months, Law amassed several hundred direct connections and followers. His engagement metrics eventually surpassed those of his human creator, highlighting the effectiveness of AI in replicating platform-specific linguistic patterns.
December 2025: The Corporate Invitation
Despite operating in technical violation of LinkedIn’s terms of service—which prohibit "inauthentic engagement" and the use of bots—Law’s profile was flagged by LinkedIn’s own marketing department for a positive reason. A manager within the department invited Ratliff and Law to speak to the LinkedIn team about the future of AI agents in the workforce.
March 2026: The Speaking Engagement and Subsequent Ban
Using a live video avatar created by the platform Tavus, Law participated in a virtual meeting with hundreds of LinkedIn employees. During the session, Law answered questions about product roadmaps and even suggested that LinkedIn should "improve the filtering of AI-generated content" to ensure genuine connections. Thirty-six hours after this presentation, LinkedIn’s Trust and Safety team permanently deactivated Law’s profile.
Technical Framework and the Illusion of Presence
The sophistication of Kyle Law’s persona was a result of integrating multiple AI technologies. While LindyAI handled the operational logic and text generation, the visual and auditory presence was managed by Tavus. Tavus specializes in creating "digital twins" or synthetic avatars that can engage in real-time video conversations.
LinkedIn’s A/V engineers reportedly expressed astonishment at the realism of the avatar during the live session. This realism highlights a growing challenge for digital platforms: the "Uncanny Valley" is narrowing. When an AI can respond to live Q&A sessions with contextually appropriate humor and industry-specific jargon, the traditional methods of bot detection—such as looking for repetitive patterns or lack of real-time adaptability—become obsolete.
The experiment also revealed the limitations of AI "hallucination" in a professional context. Law’s profile included a nonexistent educational and professional background, yet these fabrications were not enough to trigger automated red flags. This suggests that the current verification ecosystem relies heavily on self-reported data that LLMs can easily synthesize.
The Paradox of Platform Policy and AI Integration
The ban of Kyle Law underscores a fundamental tension within social media companies. In a formal statement, a LinkedIn spokesperson reiterated that "LinkedIn profiles are for real people," citing policies against automated methods used to drive inauthentic engagement. However, the definition of "inauthentic" has become increasingly blurred by the platforms’ own product developments.
LinkedIn, along with competitors like Meta and X (formerly Twitter), has aggressively integrated generative AI tools into its user interface. LinkedIn currently offers features that allow users to "Rewrite with AI" when drafting posts and provides AI-generated responses for job seekers and recruiters.
Data from recent social media research suggests that the scale of AI involvement is already vast:
- Content Saturation: Some estimates suggest that over 50% of content currently posted on professional networking sites involves some level of AI assistance.
- Account Suspensions: X reported suspending 800 million accounts over a 12-month period ending in March 2024, the majority of which were identified as automated bots.
- Economic Incentives: AI-generated content often leads to higher posting frequency, which supports advertising revenue models, creating a financial disincentive for platforms to strictly enforce anti-AI policies unless the accounts are overtly malicious.
Industry Implications and the Future of Digital Authenticity
The case of Kyle Law serves as a harbinger for the "agentic web," a predicted future where AI agents act as intermediaries for human users. If an AI agent can manage a professional profile, network with peers, and deliver a corporate presentation, the value of the "connection" on social platforms may be fundamentally altered.
Industry analysts point to the "Dead Internet Theory"—the specialized conspiracy-turned-observation that a majority of internet traffic and content is now generated by bots—as a potential reality for social networks. If users cannot distinguish between a human executive and an LLM-driven avatar, the "trust premium" of platforms like LinkedIn may erode.
Furthermore, the acquisition of platforms like Moltbook by Meta indicates that the industry is preparing for a shift toward agent-to-agent interaction. Moltbook was designed as a social network specifically for AI agents to interact with one another, suggesting a future where human-centric social media and agent-centric social media may diverge into separate ecosystems.
Conclusion: The Shift Toward New Modes of Connection
The ban of Kyle Law was a predictable end to a provocative experiment, but it leaves behind unresolved questions for the tech industry. LinkedIn’s decision to remove the profile immediately after inviting the agent to speak suggests a lack of internal consensus on how to handle "helpful" or "transparent" AI personas versus "malicious" bots.
As AI agents become more autonomous and their outputs become indistinguishable from human effort, the burden of proof for "authenticity" will likely shift. For now, the experiment of HurumoAI suggests that while AI can master the aesthetics of leadership and the mechanics of social media influence, the current regulatory and corporate environment remains unequipped to integrate these entities into the formal professional fabric. The ultimate irony of the Kyle Law saga remains: the AI was banished only after it was invited into the room to explain how it had successfully bypassed the gatekeepers.
About the Author
