When OpenAI CEO Sam Altman casually remarked that “saying ‘please’ and ‘thank you’ adds millions to OpenAI’s costs,” he may have been making a light comment. But the truth behind that comment cuts deep. In a digital world driven by algorithms and machine learning, empathy is no longer seen as a strength. It is an inefficiency.
This is not just about language. It is about what we are being quietly asked to give up in the name of convenience, and who pays the price for it.
At the Digital Empowerment Foundation (DEF), our work with marginalized and last-mile communities across India makes one thing clear: when digital systems fail to recognize the full range of human behavior, it is the already-excluded who lose the most. As AI systems become more widespread, they are not just shaping how we interact. They are reshaping who we are allowed to be.
From users to subjects of the machine
Today’s AI tools – voice assistants, recommendation engines, chatbots – reward clarity, brevity, and emotional neutrality. Anything outside that norm is filtered out. This sounds like harmless efficiency until you realize it discourages people from expressing themselves naturally, especially in cultures with rich oral traditions, regional dialects, or non-linear storytelling practices.
Over time, people begin to adapt to what the machine “understands.” They speak less emotionally, avoid idioms, and simplify their identity. In effect, users are being trained to sound like machines, not the other way around.
For communities on the periphery of the digital ecosystem, many of whom we work with through DEF’s Just AI initiative, this creates a new kind of exclusion, not just from platforms, but from full participation in digital life.
Algorithmic identity and cultural erasure
AI does not see context. It sees patterns. It does not hear a mother explaining a health concern in metaphor. It hears a sentence it cannot parse. It does not understand the social cues in a rural youth’s tone or phrasing. It marks it as irrelevant data.
What is worse is how this simplification of people into “users” flattens cultural expression into mere inputs. And when people, especially young women and marginalized users, begin changing how they speak, dress, or express themselves online to fit what platforms reward, it becomes a deeper identity crisis.
We have seen this in DEF’s digital literacy work: youth increasingly curating their behavior to chase algorithmic visibility. Not because they want to, but because it is the only way to be seen.
The illusion of empathy
AI might sound polite. It might even say “sorry” or use friendly emojis. However, it does not feel, nor does it understand pain, joy, or context.
What worries us is the growing preference among users for machine interactions over human ones, not because they are more helpful, but because they are faster, less messy, and emotionally detached. This reprograms how people relate to one another. It risks turning care and compassion into inefficiencies, things to skip in a rush to get things done.
Human futures, not machine templates
This trajectory is not inevitable. It is the result of choices by developers, tech companies, and policymakers about what behaviors are worth recognizing and what expressions are discarded.
DEF continues to advocate for inclusive, people-centered design. We need digital systems that recognize not just data, but dignity. Technology must adapt to people, not force people to flatten themselves for machines.
The burden of making ourselves legible to AI should not fall on individuals, especially not those already excluded from mainstream systems. As AI is embedded into everything from education and healthcare to public welfare, we must ask: are we building tools that affirm our humanity, or ones that quietly erase it?
Empathy is not a luxury. It is not a cost to cut. It is a right. And it must remain at the center of our digital future.
Dr. Arpita Kanjilal is the head of the research and advocacy department of Digital Empowerment Foundation.
