The New Midnight Oil
In offices and homes across India, a new work pattern is emerging. Professionals are logging back on long after dinner, not just to clear their inbox, but to collaborate with AI. They are using generative AI to draft reports, analyse data, brainstorm
ideas, or simply get a head start on the next day. Some tech professionals report spending eight to 20 hours a week outside of their jobs just to keep up with the latest AI tools. This isn't just about catching up; for many, the quiet of the night, free from meetings and interruptions, feels like the only time to truly explore what these powerful new tools can do. This after-hours usage often stems from a mix of genuine curiosity and the pressure to stay competitive in a field that is constantly evolving.
The Productivity Paradox
The promise of AI was simple: automate tedious tasks to free up human creativity and reduce workloads. The reality, however, is proving far more complex. While many users report feeling more productive, studies are beginning to reveal a troubling trend known as the 'AI Productivity Paradox'. Research has found that instead of leading to shorter workdays, AI adoption often correlates with increased work intensity. A study from UC Berkeley noted that employees using AI worked faster, took on a wider scope of tasks, and extended their work into more hours of the day, often without being asked. The time saved by AI is frequently filled with higher output expectations, additional projects, and the overhead of managing the AI tools themselves, leading to a cycle of burnout. More than half of daily AI users report feelings of burnout, a significantly higher rate than those who don't use AI.
Blurring Boundaries and Cognitive Costs
The habit of turning to AI late at night dissolves the already fragile boundaries between work and personal life. When work can be started or continued at any moment with an AI prompt, the natural stopping points of the day disappear. This 'always-on' mentality comes with hidden cognitive costs. Over-reliance on AI can lead to the atrophy of critical skills, making independent thinking and writing feel more difficult. Experts also warn of 'cognitive debt', where the constant context-switching between human thought and AI tools leads to attention depletion and mental fatigue. Furthermore, the time spent 'botsitting'—reviewing, correcting, and re-running AI outputs—is significant, with one study finding that for every hour of useful output, workers spend another hour making it usable.
The Rise of the 'Human Debt'
While companies chase the short-term gains of AI adoption, they risk accruing a 'human debt'. This debt manifests as employee burnout, higher stress levels, and a growing sense of isolation. One global survey found that while AI helped reduce some stress, 43% of respondents cited reduced human-to-human interaction as their top worry. The constant pressure to keep up, combined with unclear expectations from management, can amplify workplace anxiety. Some employees even report that AI has increased their workloads, directly contradicting its intended purpose. If the focus remains solely on maximising output, companies may see a decline in the very human qualities AI cannot replicate: judgment, creativity, and well-being.
Forging Healthier AI Habits
Navigating this new landscape requires a conscious effort from both individuals and organisations. For individuals, a key strategy is the 'attempt first' rule: try to solve a problem or draft an outline independently before turning to an AI tool. This protects and exercises core cognitive skills. For companies, the solution lies in setting clear boundaries. Leaders must shift from celebrating constant communication to protecting 'deep work' time for their teams. Establishing clear guidelines on where to use AI (for research and support) and where not to (for final decisions or core strategic thinking) is crucial. Rather than letting AI passively expand workloads, organisations need to proactively define what 'done' looks like and protect employees from an endless cycle of AI-fueled productivity.
