All questions
1959 published questions.
What happens when ai onboard a satellite has to make an emergency collision decision on its own?
When AI onboard a satellite must make an emergency collision decision on its own, it generally executes a predetermined evasive maneuver based on pre-programmed rules and real-time tracking of the approaching object, since communication delay to ground control makes real-time human decision-making impossible within the short response window.
What is a jailbreak attempt and how is it different from prompt injection?
A jailbreak attempt is a direct effort by a user to convince an AI model to bypass its own safety guidelines through clever prompting, while prompt injection instead hides malicious instructions within external content the AI processes, meaning the key distinction is whether the attack comes directly from the user's own request or is hidden within separate data the AI is asked to handle.
What is a micro credential and how does it differ from a full certification?
A micro-credential verifies a specific, relatively narrow skill, typically completed in a much shorter time than a full certification covering broader curriculum, making it well suited for quickly demonstrating a targeted skill rather than the comprehensive knowledge base a full certification represents.
What is a proof of concept and why do businesses run one before full ai adoption?
A proof of concept is a small-scale, limited trial of an AI tool conducted before committing to full organizational adoption, letting a business validate that the tool genuinely delivers expected value for its actual use case and address potential problems on a smaller, less costly scale before broader rollout.
What is a right to explanation and does it exist for ai decisions?
A right to explanation refers to a legal principle that individuals affected by an automated decision are entitled to understand its basis, and while referenced in regulations like the EU's GDPR, its actual practical scope and enforceability remain genuinely debated rather than being an unambiguous right everywhere.
What is a soft robot and how does ai control its unconventional movement?
A soft robot is built from flexible, deformable materials rather than rigid mechanical components, enabling movement closer to biological organisms like worms or octopi, and AI helps control this by learning to predict how the flexible body will deform, since traditional rigid-robot control doesn't directly translate.
What is a system card and how does it differ from a model card?
A system card documents how an AI model is actually deployed in a real product, including safety testing, guardrails, and intended use restrictions, while a model card focuses more narrowly on the underlying model itself, like its training data and technical capabilities, making the two documents complementary rather than interchangeable.
What is a wrapper startup and why do investors view them skeptically?
A wrapper startup is a company whose product largely consists of a thin interface layer over an existing foundation model's API without substantial additional differentiation, and investors view these skeptically since they risk easy replication by competitors or redundancy if the underlying provider adds a similar feature directly.
What is an ai bill of rights and has any government actually adopted one?
An AI bill of rights refers to a proposed set of principles outlining protections individuals should have regarding AI systems affecting them, including a notable non-binding blueprint published by the White House, and while these articulate important principles, they generally function as non-binding guidance rather than enforceable law.
What is an ai center of excellence and why do larger companies create one?
An AI center of excellence is a centralized internal team responsible for coordinating AI strategy, sharing best practices, and providing specialized expertise across an organization, and larger companies create one to avoid duplicated effort and inconsistent practices when departments pursue AI adoption independently.
What is an ai incident database and why do researchers maintain one?
An AI incident database is a maintained collection of documented cases where an AI system caused harm or behaved in an unintended way, and researchers maintain these to help the field learn from real-world failures, identify recurring patterns across systems, and inform better safety practices rather than repeating past mistakes.
What is an ai models parameter count and why does it matter less than it used to?
An AI model's parameter count is the number of adjustable internal values that determine its learned behavior, and while a larger count historically correlated with greater capability, this relationship has weakened as more efficient training techniques and better data let smaller models match or exceed older, larger ones.
What is catastrophic forgetting in ai training?
Catastrophic forgetting refers to an AI model losing previously learned capability when it's trained on new information, since the same internal parameters used to store earlier knowledge get overwritten during subsequent training, making this a genuine risk companies have to actively manage whenever fine-tuning a model for a new specific task.
What is data exfiltration risk in ai connected browser agents?
AI-connected browser agents, which can autonomously navigate websites and take actions on a user's behalf, carry genuine data exfiltration risk since a malicious website could potentially manipulate the agent through embedded hidden instructions into revealing sensitive information it has access to, or taking unintended actions, a risk that grows as these agents are granted broader system access.
What is dual use risk in the context of ai safety policy?
Dual-use risk in AI safety policy refers to the reality that many AI capabilities genuinely useful for legitimate purposes can also be misused for harm, like AI research accelerating drug discovery also potentially informing harmful biological agent design, creating a genuine challenge in governing capabilities that are simultaneously valuable and dangerous.
What is geofencing and how does it limit where autonomous vehicles can operate?
Geofencing defines a specific geographic boundary within which an autonomous vehicle is permitted to operate, restricting it from attempting full autonomy outside this predetermined, thoroughly tested area, reflecting a deliberate safety approach limiting deployment to conditions the system has actually been validated for.
What is model watermarking and can it help trace leaked ai outputs?
Model watermarking embeds a subtle, statistically detectable pattern into an AI model's generated output that doesn't affect normal quality but can later be identified using a specific detection method, helping trace whether a specific piece of content actually originated from that model, though watermarks can sometimes be removed or degraded through subsequent editing of the output.
What is parametric insurance and how does ai make it possible?
Parametric insurance pays out automatically based on a predefined measurable trigger, like a hurricane reaching a specific wind speed, rather than a traditional damage assessment process, and AI makes this increasingly practical by rapidly analyzing real-time weather and sensor data to confirm when a trigger has genuinely been met.
What is speculative decoding and how does it speed up ai responses?
Speculative decoding speeds up AI responses by using a smaller, faster draft model to quickly guess several upcoming words at once, which the larger, more capable model then verifies in a single pass rather than generating each word one at a time, producing the same quality output considerably faster when the draft guesses turn out to be correct.
What is test time compute and how does it improve ai reasoning?
Test-time compute refers to the additional computational effort an AI model spends actually working through a problem at the moment it's asked, rather than during its original training, and increasing this effort — letting a model reason through more intermediate steps before answering — has proven to genuinely improve accuracy on complex reasoning tasks beyond what training alone achieves.
What is the difference between a foundation model and a fine tuned model?
A foundation model is a large, general-purpose AI model trained on broad data to develop wide-ranging capability, while a fine-tuned model starts from that same foundation but undergoes additional, more targeted training on specific data to specialize its behavior for a narrower task, making fine-tuning a refinement step built on top of a foundation model rather than a separate starting point.
What is the difference between ai safety research and ai capabilities research?
AI safety research focuses on ensuring AI systems behave reliably and in line with human intentions, while AI capabilities research focuses on expanding what AI systems can actually do, and while conceptually distinct, these two areas are genuinely interconnected since more capable models often need more sophisticated safety measures.
What is the risk of ai tools reinforcing existing inequalities in aid distribution?
AI tools used in aid distribution carry a genuine risk of reinforcing existing inequalities if the historical data they're trained on reflects past patterns where certain groups had less access to registration systems, potentially causing an AI-driven system to systematically underserve populations already historically underserved.
What is the risk of vendor lock in with a single ai platform provider?
The risk of vendor lock-in with a single AI platform provider is that a business becomes so deeply integrated with that provider's particular features that switching later becomes genuinely difficult and costly, leaving limited negotiating leverage if that provider's pricing, terms, or service quality change unfavorably.