All questions
1959 published questions.
What are the privacy implications of AI based employee monitoring software?
AI-based employee monitoring software raises significant privacy concerns, including detailed behavioral data collection that can extend into personal time or devices, uncertainty about data retention and access, and effects of constant surveillance on trust and wellbeing, with legal protections varying by jurisdiction.
What are the risks of using AI to make decisions about who receives aid?
Using AI to decide who receives humanitarian aid carries genuine risks, including biased or incomplete data producing unfair outcomes, reduced human judgment in decisions with life-affecting consequences, and accountability gaps when something goes wrong, which is why responsible organizations maintain human oversight.
What caused the first AI winter?
The first AI winter, occurring roughly in the mid-to-late 1970s, was caused primarily by a combination of overpromised research results failing to materialize and influential critical government reports — including the UK's Lighthill Report and the US ALPAC report on machine translation — that led major funding agencies to sharply cut back research support after early optimism proved premature.
What data do AI yield prediction models actually rely on?
AI yield prediction models typically rely on a combination of historical yield records, real-time weather data, soil condition information, satellite or drone imagery showing current crop health, and sometimes management practice data like planting date and input application, combining these sources to identify patterns associated with different yield outcomes.
What data privacy concerns come up when farms use AI monitoring tools?
Common data privacy concerns with farm AI monitoring tools include uncertainty about who owns and can access collected farm data, whether that data could be shared with or sold to third parties like input suppliers or insurers without clear farmer consent, and how securely sensitive operational data is stored and protected against breaches.
What does a prompt engineer actually do day to day?
A prompt engineer's day-to-day work typically involves designing, testing, and refining instructions that get reliable behavior out of a large language model, plus building evaluations to measure whether changes actually improve output quality — though as a stand-alone title it's become less common than in the field's early days.
What does an AI safety job actually involve?
AI safety roles generally involve identifying and reducing risks from AI systems — through technical work like alignment research and red-teaming, or through policy and governance work like drafting usage guidelines and risk frameworks — with the exact mix of technical versus policy focus varying significantly by role and organization.
What does narrow AI versus general AI actually mean?
Narrow AI refers to systems built to perform one specific task or a limited set of related tasks well, which describes essentially all AI systems in use today, while general AI (AGI) refers to a hypothetical system with broad, human-comparable intelligence across many tasks — something that does not currently exist.
What does training a model actually mean at a basic level?
At a basic level, training a model means repeatedly showing it examples, comparing its output against a known correct answer or a defined measure of quality, and automatically adjusting its internal parameters a small amount each time to reduce the gap between its output and the desired result, until performance stabilizes at an acceptable level.
What ended the most recent AI winter and started the current boom?
The most recent AI winter gradually ended through the 2000s and early 2010s as growing computing power, larger datasets, and neural network refinements accumulated, culminating in the visible 2012 ImageNet deep learning breakthrough, widely credited with convincing the field and funders a sustained period of progress had begun.
What free resources do practitioners actually recommend for learning AI?
Practitioners commonly recommend official documentation from AI tool providers, freely available university course material, hands-on practice with widely used AI tools directly, and active participation in practitioner communities as some of the highest-value free ways to build genuine AI skills.
What happens if AI wrongly flags a legitimate claim as fraudulent?
When AI wrongly flags a legitimate claim as fraudulent, well-designed insurer processes route it to a human fraud investigator rather than automatically denying it, so the policyholder generally experiences delay and scrutiny during investigation, but the claim should be processed normally once confirmed legitimate.
What happens when a nonprofit can't afford to maintain an AI tool after a grant ends?
When a nonprofit can't afford to maintain an AI tool after grant funding ends, it generally faces difficult choices: discontinuing the tool and reverting to prior processes, seeking additional sustaining funding, or scaling back use — a common, documented challenge in nonprofit technology adoption.
What happens when a self-driving car breaks a traffic law?
When a self-driving car breaks a traffic law, consequences and responsibility depend on autonomy level and jurisdiction, with the human occupant generally still held responsible at lower levels, while higher-autonomy situations raise novel questions for enforcement designed around human drivers.
What is a neural network explained without the jargon?
A neural network is a computing system loosely inspired by how brain cells connect, made up of many simple processing units organized in layers, where each connection has an adjustable 'weight' that the system tunes during training to gradually get better at turning a given input into a correct or useful output.
What is a prompt injection attack and why does it matter?
A prompt injection attack involves inserting malicious instructions into content an AI system processes — like a document or webpage it's asked to summarize — to hijack its behavior toward the attacker's hidden instructions, and it matters because it can cause data leaks, unintended actions, or harmful output.
What is an adversarial attack on an AI model?
An adversarial attack on an AI model is a deliberate attempt to manipulate its behavior or output by feeding it specially crafted input designed to exploit weaknesses in how it processes information, ranging from subtly altered images causing misclassification to prompts bypassing a language model's intended restrictions.
What is an AI product manager responsible for?
An AI product manager is responsible for deciding what an AI-powered product should do and for whom, translating between research/engineering teams and end users, setting quality and safety bars for model behavior, and prioritizing tradeoffs unique to AI products like reliability, latency, and cost per query.
What is an algorithmic impact assessment and when is one required?
An algorithmic impact assessment is a structured evaluation, conducted before or during an AI system's deployment, examining its potential effects on individuals — including bias, privacy, and accuracy risks — increasingly required for higher-risk government AI use cases under evolving policy, though not yet universal.
What is data poisoning and how does it compromise an AI model?
Data poisoning is an attack technique where an attacker deliberately introduces manipulated data into a model's training dataset to compromise its resulting behavior — causing misclassification, a hidden exploitable vulnerability, or biased output — posing particular risk for models trained on unverified scraped data.
What is New York City's Local Law 144 and why does it matter for AI hiring tools?
New York City's Local Law 144 regulates automated employment decision tools by requiring covered employers to conduct independent bias audits, publish the results, and notify candidates when such tools are used — making it one of the most prominent examples of AI hiring regulation in the U.S.
What is proxy discrimination and why does it matter for insurance AI?
Proxy discrimination occurs when a seemingly neutral factor in an insurance AI model closely correlates with a protected characteristic like race, producing discriminatory outcomes even without directly using that characteristic — a significant concern since sophisticated models can find many such subtle correlations.
What laws currently govern how the US federal government can use AI?
U.S. federal government AI use is currently governed by a combination of executive orders, agency-specific guidance from bodies like OMB, voluntary frameworks like NIST's AI Risk Management Framework, and existing privacy, civil rights, and administrative law — rather than a single, comprehensive federal AI-specific statute.
What laws currently regulate AI use in hiring decisions?
AI use in hiring decisions in the U.S. is currently regulated through existing federal anti-discrimination law that applies regardless of AI involvement, specific EEOC guidance for automated tools, and a growing number of state and local laws — like NYC's Local Law 144 — requiring bias audits and disclosure.