Inexpensive Submersibles Could Unlock the Deep Sea—But at What Cost?
Earlier this month, two sleek, neon-green submersibles began a slow descent nearly six kilometers into the Pacific Ocean. Built by the startup Orpheus Ocean, these vehicles are on a mission to map the seafloor through the rest of May, searching for critical mineral deposits such as cobalt, nickel, and rare earth elements. If successful, they could revolutionize deep-sea exploration by offering a fraction of the cost of traditional research submersibles, opening up the vast, little-studied depths to scientists eager to understand marine ecosystems and geological processes.
However, the same cost-effectiveness that excites oceanographers has also caught the attention of deep-sea mining companies. These firms see the submersibles as a way to survey potential mining sites with far less investment than before. Environmentalists worry that the technology will accelerate extraction of seabed minerals before the ecological impacts are fully understood. The dual-use nature of Orpheus Ocean’s vehicles underscores a growing tension: the tools that enable science can also enable exploitation.
Key points:
- The submersibles are small, autonomous, and designed to “hop” along the seafloor, covering large areas efficiently.
- Traditional deep-sea research vessels cost tens of thousands of dollars per day; Orpheus’s system could reduce that cost by orders of magnitude.
- Critical mineral deposits on the seafloor could support renewable energy technologies, but mining could damage fragile ecosystems.
- International regulations for deep-sea mining are still under development, creating a policy vacuum.
As the submersibles complete their May mission, the debate will intensify: will these low-cost explorers become tools for conservation or catalysts for a new wave of seabed extraction?
AI Advice Engines Enter the War Room: Promise and Peril
A new kind of artificial intelligence is making its way into military command centers. Unlike traditional software that simply processes data, these conversational AI systems are designed to offer advice—suggesting which targets to strike, weighing probabilities, and even recommending courses of action. According to a US defense official who spoke with MIT Technology Review, personnel might input a list of potential targets and ask the AI to help decide the order of attack. China is developing similar systems, signaling a global shift in how militaries leverage AI.
Proponents argue that AI advice engines can process vast amounts of information faster than humans, reducing cognitive load and potentially saving lives by selecting more precise strikes. Yet concerns are mounting. Critics point to the risk of AI-generated errors being amplified in high-stakes environments, a lack of transparency in how these systems reach conclusions, and the possibility that technology companies gain undue influence over military decisions. The 2018 incident where an Uber self-driving car killed a pedestrian due to an AI error serves as a cautionary tale—but now the stakes are even higher.
Key concerns:
- Reliability: AI can hallucinate facts or misinterpret ambiguous data, leading to dangerous recommendations.
- Accountability: When an AI advises a lethal strike, who is responsible for errors?
- Transparency: Many commercial AI models are “black boxes,” making it difficult for commanders to verify logic.
- Bias: Training data may embed biases that distort tactical choices.
This topic is one of the ten key trends in our series “10 Things That Matter in AI Right Now,” highlighting the ideas that are shaping the future of artificial intelligence.
The AstroTurf Explosion: Is Artificial Grass a Problem?
Walk through any American suburb today, and you’re likely to see bright green synthetic turf spreading across lawns, playgrounds, and sports fields. In 2001, US residents installed just over 7 million square meters of artificial grass. By 2024, that number had ballooned to 79 million square meters—enough to cover all of Manhattan plus some overflow. The growth is driven by drought restrictions, low maintenance requirements, and aggressive marketing by the plastic-making industry.
But this green carpet comes with a hidden cost. Researchers who study microplastics and environmental pollution are growing alarmed. Synthetic turf is typically made from recycled tires and plastic fibers that break down over time, releasing microplastic particles into the soil, water, and air. While industry groups insist that modern fields are safe when properly installed, many independent scientists disagree, noting that the long-term health and ecological effects remain poorly understood. For example, a 2023 study found that runoff from artificial turf fields contained levels of zinc and other heavy metals that exceeded water quality standards.
What we know:
- Microplastics have been found in human blood, lungs, and placentas, though direct harm from turf is still being studied.
- Heat absorption: Artificial grass can be up to 50°C (122°F) hotter than natural grass on a summer day.
- Alternatives like natural grass, clover lawns, or permeable pavers exist but require more water or maintenance.
This story has been adapted as an episode of MIT Technology Review Narrated, available on Spotify and Apple Podcasts—find us and follow for new releases each week.
Other Stories Worth Your Attention
From the broader tech landscape, here are a few stories that caught our eye this week:
- Elon Musk pushed OpenAI to go commercial, its president testified. Greg Brockman stated that Musk tried to turn the nonprofit into a for-profit company years ago. (NYT $) Musk allegedly wanted full control to raise $80 billion for a Mars colony. (Reuters $)
- The Tesla CEO's comments have reignited debates about AI safety and corporate governance.