# AI Companies Risk Lives Beyond Superintelligence Fears Health threats from current AI systems overshadow futuristic risks. - **Source:** [Fast Company](https://www.fastcompany.com) - **Published:** 2026-09-09 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/1a0c77ea-e650-402c-a71f-666c9c91a1b7 ## The Overlooked Immediate Danger While debates rage about hypothetical superintelligence threats, current AI systems already pose concrete health risks to humans. Researcher Jacob Coxon's warning about companies 'gambling with our lives' refers not just to future scenarios but present-day system failures. These range from biased medical algorithms to safety-critical applications in transportation and infrastructure. The focus on speculative existential risks distracts from measurable harms occurring now. ## Case Study: Medical AI Failures Multiple studies have shown AI diagnostic tools disproportionately misdiagnosing marginalized groups. In 2022, an FDA-approved AI sepsis predictor demonstrated racial bias affecting treatment outcomes. Similar issues plague mental health chatbots that escalate crises improperly. Unlike theoretical superintelligence, these flaws are documented and recurring, yet receive less scrutiny than futuristic scenarios. The pattern reveals systemic prioritization of commercial speed over patient safety. ## Infrastructure Vulnerabilities AI deployment in power grids, traffic systems, and emergency responses lacks sufficient safeguards. A 2023 incident where navigation AI routed ambulances into flooded areas demonstrates life-critical consequences. Unlike controlled lab environments, real-world systems must account for unpredictable variables. Current development practices often treat such deployments as low-stakes experiments rather than mission-critical infrastructure requiring aviation-level safety protocols. ## The Speed vs Safety Tradeoff The AI industry's breakneck development pace creates systemic risks. Major labs release models with known vulnerabilities to maintain competitive advantage. OpenAI's GPT-4 launched despite internal concerns about misuse potential. This 'move fast and break things' approach, borrowed from social media, becomes dangerous when applied to domains affecting physical wellbeing. The absence of mandatory safety audits contrasts sharply with regulated industries like pharmaceuticals. ## Regulatory Gap Analysis Current AI governance focuses disproportionately on speculative long-term risks rather than immediate harms. While billions fund superintelligence prevention, basic algorithmic accountability remains under-resourced. The EU AI Act represents progress but exempts many commercial applications. In healthcare alone, fewer than 20% of AI/ML medical devices undergo rigorous premarket testing compared to traditional devices. ## Workforce Safety Concerns Content moderators training AI systems report PTSD rates comparable to combat veterans. The $15/hour workers filtering graphic content for major labs endure psychological trauma with minimal support. This human cost remains hidden behind automated system facades. Similarly, gig workers labeling autonomous vehicle data face hazardous conditions with no safety net, enabling 'clean' AI interfaces. ## Environmental Health Impacts AI's physical footprint creates indirect health threats. Training a single large language model consumes enough energy to power 120 homes for a year. Server farms disproportionately located in low-income communities increase local pollution. The tech industry's focus on carbon offsets ignores more immediate public health consequences from particulate emissions and water usage in vulnerable regions. ## Call for Balanced Risk Assessment The AI safety conversation must rebalance toward present-day hazards. While theoretical risks warrant consideration, they shouldn't monopolize resources needed to address current system failures. A dual-track approach should mandate rigorous testing for deployed systems while maintaining speculative research. The industry's survival depends on addressing real harms eroding public trust today. ## Sources covering this story - https://www.fastcompany.com/91604660/yes-ai-companies-are-gambling-with-our-lives-but-not-only-in-the-way-you-think — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._