Truth 1: AI is hungry for electricity. In order to train and operate large AI models, huge data centers are required, which use a growing amount of power. By 2030, according to recent studies, the amount of electricity used in data centers around the world could roughly double, with AI-specific servers using tens of terawatt hours per year. This would power millions of homes and put a real strain on grids and climate goals.
Truth 2: AI quietly shapes who gets opportunities. AI systems used in hiring, lending, policing, healthcare, and advertising can reproduce and amplify human bias because they learn from biased historical data. Cases where algorithmic decisions have discriminated against people on the basis of race, gender, or other sensitive characteristics have been documented in studies, transforming seemingly invisible math into very real discrimination in employment, credit, and medical care.
Truth 3: AI will rewrite jobs, not just “assist.” Economic research suggests that a large share of jobs in advanced and emerging economies are exposed to AI, especially roles with routine cognitive tasks like basic analysis, customer support, and data entry. Some workers and sectors will gain productivity and new roles, but others may see outright displacement or slower recovery after recessions, making digital and AI skills a survival requirement rather than a luxury.
The fourth truth is that very few hands have control. Due to the high costs of data, chips, energy, and talent, only a small number of businesses and governments can afford to build and train frontier-scale AI models. Because of this concentration, decisions regarding access rules, safety limits, data use, and model behavior are actually decisions regarding information power and economic advantage for entire societies.
Truth 5: “Neutral” AI always reflects human values. Because humans choose the data, goals, and safety guidelines, there is no such thing as an AI model that is completely impartial. Since every system encodes trade-offs between accuracy, profit, fairness, privacy, and control, the true unanswered question of any AI is whose interests and values it is optimized to serve?
A concise, thought-provoking article that exposes AI’s environmental costs, social biases, labor disruption, power concentration, and value-laden design, compelling readers to question who truly controls AI’s growing influence.
You must be logged in to post a comment.