A
Ali Kolahdoozan
Guest
More often indirectly β By framing debates in ways that avoid scrutiny. Instead of asking βWhich AI, for which problem, at what cost?β the discussion becomes βYouβre either with us, or youβre left behind.β
By omission β Leaving out ecological and human costs, or burying them in footnotes and appendices.
What Real Transparency Would Look Like
If CEOs were truly honest, we would see:
Resource accounting β Exact electricity, water, and carbon use for each product and training run.
Data provenance β Where data comes from, under what consent, and how people can opt out.
Labor visibility β Who is labeling and moderating, under what conditions, and at what pay.
Impact metrics β Beyond benchmarks, measuring actual problem-solving and social effects.
Community rights β Local communities hosting infrastructure should have real veto power.
What Can We Do?
Demand details, not slogans: ask βWhich model, which data, what cost?β
Shape policies in schools, workplaces, and governments to set responsible AI standards.
Defend data rights: insist on informed consent and fair use of personal content.
Push for transparent reporting on energy, water, and labor.
Support smaller, efficient, local AI solutions over defaulting to mega-models.
Conclusion
Tech CEOs may not always βlieβ outright, but their skillful framing, selective disclosures, and sweeping promises create a distorted picture. The narrative of βinevitable AI progressβ only continues if we accept it passively.
The real question isnβt whether CEOs lieβitβs: how do we build systems where, even if they wanted to, they couldnβt?
Continue reading...
By omission β Leaving out ecological and human costs, or burying them in footnotes and appendices.
What Real Transparency Would Look Like
If CEOs were truly honest, we would see:
Resource accounting β Exact electricity, water, and carbon use for each product and training run.
Data provenance β Where data comes from, under what consent, and how people can opt out.
Labor visibility β Who is labeling and moderating, under what conditions, and at what pay.
Impact metrics β Beyond benchmarks, measuring actual problem-solving and social effects.
Community rights β Local communities hosting infrastructure should have real veto power.
What Can We Do?
Demand details, not slogans: ask βWhich model, which data, what cost?β
Shape policies in schools, workplaces, and governments to set responsible AI standards.
Defend data rights: insist on informed consent and fair use of personal content.
Push for transparent reporting on energy, water, and labor.
Support smaller, efficient, local AI solutions over defaulting to mega-models.
Conclusion
Tech CEOs may not always βlieβ outright, but their skillful framing, selective disclosures, and sweeping promises create a distorted picture. The narrative of βinevitable AI progressβ only continues if we accept it passively.
The real question isnβt whether CEOs lieβitβs: how do we build systems where, even if they wanted to, they couldnβt?
Continue reading...
