The technological landscape of late May 2026 presents a stark dichotomy. On one side, venture capital continues to flow into artificial intelligence at an unprecedented rate, fueling breakthroughs in autonomous agents and specialized processors. On the other, a growing unease pervades the workforce and regulatory bodies, grappling with the ethical fallout of AI’s rapid deployment. This tension, which Box founder Aaron Levie recently termed “AI psychosis,” encapsulates the dilemma: a fervent belief in AI’s transformative power often blinds decision-makers to the complexities of human work and the fundamental right to privacy.
The Automation Imperative and the Shifting Workforce
The promise of AI has always been rooted in efficiency and augmentation, but its immediate impact on the job market has been far more disruptive. Companies, eager to capitalize on perceived cost savings and productivity gains, are increasingly turning to AI agents to shoulder tasks historically performed by humans. A recent case in point is
, a project management software company, which reportedly cut 22% of its workforce, attributing the decision to the adoption of AI agents. This move, while perhaps driven by a pursuit of operational streamlining, casts a long shadow over the future of work.
This isn’t an isolated incident, but rather a symptom of a broader trend where tech layoffs in 2026 are already approaching the cumulative figures for all of 2025. The logic is compelling on a spreadsheet: AI agents can operate 24/7, scale instantly, and theoretically perform repetitive tasks with fewer errors. Yet, Levie’s observation rings true: the individuals championing these AI-driven workforce reductions are often the furthest removed from the intricacies of the jobs they aim to replace. They may see a task as a series of inputs and outputs, overlooking the tacit knowledge, emotional intelligence, and complex problem-solving that define many human roles.
The question then becomes: are these AI agents truly replacing, or simply augmenting? The distinction is crucial. Augmentation implies empowering humans to achieve more, while replacement suggests an outright substitution. The current wave of AI adoption often leans towards the latter, particularly in areas like customer service, content generation, and even coding.
The Double-Edged Sword of AI Coding Agents
Nowhere is this ambiguity more evident than in the realm of software development.
, a two-year-old startup, made significant waves by raising $1 billion at a staggering $26 billion valuation this year, all on the strength of its flagship product, Devin. Touted as the first and arguably most successful AI coding agent, Devin is designed to “naturally own tasks end to end,” according to Cognition CEO Scott Wu. The company’s vision, articulated in its announcement of the funding round, paints a picture of “self-driving software development,” where autonomous systems handle complex coding projects from conception to deployment.
Wu, despite the ambitious rhetoric, maintains that Devin is not designed to supplant human programmers. “We’ve never thought about it as replacing humans,” he stated, countering the widespread anxiety about AI leading to mass unemployment in the tech sector. He suggests that Devin acts more as a “co-worker,” freeing human developers from mundane tasks and allowing them to focus on higher-level problem-solving and innovation. This perspective, however, requires careful scrutiny. If Devin can indeed “own tasks end to end,” the line between augmentation and replacement becomes exceptionally thin for certain roles, particularly those that involve routine coding, debugging, and integration. It points to a future where human programmers may need to evolve their skill sets dramatically, moving from code generation to AI oversight and strategic architecture.
This discussion also brings into focus the evolving definition of Artificial General Intelligence (AGI). While still a nebulous term, AGI generally refers to AI systems capable of performing most human intellectual tasks at or above the average human level. Companies like OpenAI have even defined it as an “equivalent of a median human that you could hire as a co-worker.” Devin’s capabilities, if they live up to the hype, push the boundaries closer to this ideal, raising legitimate questions about the long-term impact on the human workforce.
The Unseen Costs: Data, Privacy, and Trust
Beyond job displacement, the ethical landscape of AI is fraught with challenges related to data collection and privacy. Training powerful AI models, particularly large language models (LLMs) and advanced robotic systems, demands vast quantities of real-world data. This insatiable hunger for data often leads companies to push the boundaries of what is considered acceptable, if not legal.
Consider the case of
Model Capability Initiative (MCI). This internal AI training tool is reportedly capturing extensive computer usage data from U.S. employees, and critically, it’s also collecting non-U.S. data transmitted through these communications. This expansive data collection strategy is on a direct collision course with Europe’s stringent General Data Protection Regulation (GDPR) and other global privacy frameworks. Employees have voiced significant concerns regarding the tool’s scope, its impact on their internet usage, and the potential exposure of sensitive information. The precedent set by such internal monitoring, even under the guise of AI development, erodes trust and raises fundamental questions about employee rights in an increasingly data-driven workplace.
The need for real-world data extends beyond enterprise environments. Startups developing AI for physical tasks, such as robotics, are resorting to unconventional methods to gather training material.
, an AI training startup, for instance, offers free home cleaning services in cities like New York and London. The catch? In exchange for the cleaning, Shift demands extensive video footage of its cleaners at work, capturing every scrub, wipe, and mop. This data is invaluable for teaching robots to understand the complexities of physical spaces, object manipulation, and human interaction in unstructured environments. While seemingly benign, such practices highlight the pervasive nature of data extraction for AI development and the trade-offs individuals are increasingly asked to make.
India’s AI Ambition and the Global Context
For India, a nation rapidly positioning itself as a global technology hub, these developments present both immense opportunities and significant challenges. India’s burgeoning deep tech research ecosystem, its aggressive push for semiconductor manufacturing capabilities, and its vibrant SaaS platform sector are all heavily reliant on AI innovation. The country’s vision for EV adoption and sustainability initiatives also requires sophisticated AI for optimization and management.
As Indian enterprises and startups increasingly embrace AI to enhance competitiveness, they must navigate the same ethical tightrope. The potential for AI to drive economic growth, create new industries, and solve complex societal problems (from healthcare to climate change) is undeniable. However, the lessons from global trends, particularly concerning job displacement and data privacy, must be heeded. India’s large workforce, while adaptable, requires robust upskilling initiatives and a proactive approach to managing the transition. Furthermore, as Indian companies develop and deploy AI systems globally, they will face the same stringent regulatory scrutiny as their Western counterparts, particularly concerning data governance and algorithmic fairness. The need for a comprehensive, forward-looking AI policy that balances innovation with social responsibility is paramount. This includes fostering domestic research in ethical AI, developing explainable AI systems, and establishing clear guidelines for data collection and usage within the Indian context.
Towards a Responsible AI Future
The current “AI psychosis” is a call to action. The relentless pursuit of AI-driven efficiency, while understandable from a business perspective, must be tempered with a profound understanding of its human and ethical implications. The narratives around AI agents like Devin, which promise “self-driving software development” while simultaneously denying human replacement, illustrate the fine line between augmentation and displacement that policymakers, business leaders, and employees must navigate.
The immense investments flowing into AI, exemplified by Groq’s reported $650 million raise to scale its AI inference capabilities, underscore the technology’s foundational importance. However, this investment must be paired with an equally robust commitment to ethical deployment. The controversies surrounding Meta’s data collection practices and Shift’s real-world data acquisition highlight the urgent need for transparent data governance, stronger privacy protections, and a clearer articulation of what constitutes ethical data sourcing for AI training.
Ultimately, the future of AI will not be determined solely by its technical prowess, but by our collective ability to foster a responsible ecosystem. This means moving beyond the hype to critically assess AI’s true impact, challenging claims of pure augmentation when replacement is on the horizon, and rigorously upholding privacy rights in an increasingly data-hungry world. Only then can we build AI systems that truly serve humanity, rather than merely automating its present or displacing its future.
