Risks Associated With AI
AI comes with several major risks.[:cite[2]{ln=8}] [:cite[1]{ln=1}] Job disruption in some sectors: AI is not yet replacing large numbers of jobs in most developing countries, but it is already putting pressure on...
AI comes with several major risks.[:cite[2]{ln=8}] [:cite[1]{ln=1}] Job disruption in some sectors: AI is not yet replacing large numbers of jobs in most developing countries, but it is already putting pressure on industries such as call centers and back office services where automation can substitute for human work.[:cite[2]{ln=1}] [:cite[2]{ln=2}] [:cite[2]{ln=3}] Bias and exclusion: AI systems can be biased when they are trained on data that do not reflect local people, languages, and cultures.[:cite[2]{ln=4}] Misinformation, fraud, and cybercrime: AI can make it easier to spread false information, commit fraud, and carry out cybercrime.[:cite[2]{ln=5}] Privacy, discrimination, and security harms: Poorly designed AI systems can discriminate against people, violate privacy, and create cybersecurity risks.[:cite[1]{ln=2}] [:cite[3]{ln=5}] Dependence on a few powerful firms and countries: Advanced AI models, chips, and data centers are controlled by a small number of companies and economies, which creates dependency risks for countries that do not control the technology they use.[:cite[2]{ln=6}] [:cite[2]{ln=7}] [:cite[4]{ln=4}] Surveillance and propaganda: AI can enable governments or other actors to spread propaganda and monitor citizens, especially where institutions are weak.[:cite[4]{ln=7}] [:cite[5]{ln=1}] [:cite[5]{ln=2}] Weak accountability and “black box” decisions: Many AI tools are hard to interpret, and when they cause harm it can be unclear whether responsibility lies with the developer, vendor, agency, or user.[:cite[6]{ln=2}] [:cite[6]{ln=3}] [:cite[7]{ln=1}] [:cite[7]{ln=3}] Misuse in government: In public administration, AI can be used for deliberate manipulation or mass surveillance, and in weak institutional settings this can further erode legitimacy and make legal redress harder.[:cite[8]{ln=1}] [:cite[8]{ln=4}] Hallucinations and unsafe advice: Generative AI can produce plausible sounding but false answers, and in settings like agricultural advice this can create financial, environmental, and health risks.[:cite[9]{ln=1}] [:cite[9]{ln=2}] [:cite[9]{ln=3}] [:cite[9]{ln=4}] Environmental damage: AI has significant environmental costs across its life cycle, including mineral extraction, energy use, water use, and growing e waste, with developing countries especially vulnerable to these harms.[:cite[10]{ln=1}] [:cite[10]{ln=2}] [:cite[12]{ln=1}] [:cite[11]{ln=1}] [:cite[13]{ln=3}] Loss of trust: Failures in high stakes uses of AI can quickly erode public trust, and once trust collapses it can slow adoption and reduce long term benefits.[:cite[1]{ln=3}] [:cite[15]{ln=1}] [:cite[15]{ln=2}] [:cite[14]{ln=1}] [:cite[14]{ln=2}] ==In short, the main risks of AI are not just technical errors, but also social, economic, political, and environmental harms if AI is deployed without safeguards, accountability, and trustworthy governance.[:cite[1]{ln=5}] [:cite[16]{ln=1}] [:cite[16]{ln=2}] [:cite[16]{ln=4}]==