All questions
1778 published questions.
Can AI underwriting models use data sources beyond traditional risk factors?
Yes — AI underwriting models can incorporate a wider range of data beyond traditional risk factors, including wearable device data and vehicle telematics, though specific data sources permitted vary by insurance line and jurisdiction, since regulators require data used to have a legitimate connection to risk.
Can AI wrongly deny someone public benefits and what recourse exists?
Yes, AI systems used in benefits determinations can produce wrongful denials, whether due to data errors, flawed model logic, or edge cases the system wasn't designed to handle well, and recourse generally includes the right to request human review, appeal the decision through an agency's formal appeals process, and in some cases pursue legal action if due process rights were violated.
Can an employer be sued for using biased AI hiring software?
Yes — employers can be sued and held legally liable for using AI hiring software that produces discriminatory outcomes, since existing anti-discrimination laws apply to hiring decisions regardless of method, meaning an employer can't avoid liability by attributing an outcome to an automated tool.
Can attackers steal a proprietary AI model just by querying it?
Yes, to a meaningful degree — a technique called model extraction lets an attacker approximate a proprietary AI model's behavior by systematically querying it and training a separate model on the responses, without needing access to the original code or training data, posing a real competitive risk.
Can autonomous delivery robots and drones actually replace human delivery workers?
Not entirely, at least not yet — autonomous delivery robots and drones can handle specific, well-defined scenarios like short-distance sidewalk delivery, but genuine limitations around complex environments, diverse package situations, and varying local regulations mean human workers remain essential.
Can candidates challenge or appeal an AI driven hiring rejection?
In most cases, candidates can request reconsideration of an AI-driven hiring rejection by directly contacting the employer, though there's no universal legal right to a formal appeal, and some specific regulations, like NYC's Local Law 144, have introduced disclosure requirements giving candidates some awareness.
Can citizens find out when an AI system was used to make a decision about them?
It depends significantly on the specific agency and program — some government AI use cases require disclosure to affected individuals as part of due process protections or specific agency policy, while in other cases this disclosure isn't automatically or clearly provided, meaning citizens sometimes need to specifically request an explanation of how a decision affecting them was made.
Can insurance AI models be audited for bias?
Yes — insurance AI models can be audited for bias by analyzing outcomes across demographic groups for disparities and examining specific inputs for proxy discrimination effects, and a growing number of states require this auditing, though auditing complex models is more technically challenging than simpler ones.
Can policyholders appeal an AI driven claim denial?
Yes — policyholders generally retain the right to appeal an insurance claim denial regardless of whether AI was involved, since existing regulatory frameworks generally require insurers to provide a formal appeals process, though whether a policyholder is told AI contributed varies by insurer and jurisdiction.
Can small changes to an image really fool an AI system?
Yes — small, carefully calculated changes to an image, often imperceptible to a human observer, can genuinely cause an AI image recognition system to confidently misclassify it entirely, a well-documented phenomenon called adversarial examples, demonstrated across numerous real systems in controlled research.
Can small nonprofits actually afford to use AI tools?
Yes, many small nonprofits can afford at least some AI tools, since a growing number of providers offer free or discounted nonprofit pricing, and many general-purpose AI tools have low or no direct cost — though more specialized, comprehensive nonprofit AI platforms can still represent a meaningful cost barrier.
Can you get an AI job without a computer science degree?
Yes — many AI-adjacent roles (product, data annotation, AI-assisted operations, prompt design, technical writing, and program management) hire people without a CS degree, though core machine learning engineering roles still usually expect strong applied math or software skills, whether earned in school or on the job.
Can you learn AI skills for free or is paid training necessary?
Yes, genuinely useful AI skills can be learned for free through high-quality open courses, official documentation, and hands-on practice, though paid training can add value through structured accountability, mentorship, and credentialing that free resources typically don't provide on their own.
Did AI research really start in the 1950s or does its history go back further?
While 'artificial intelligence' as a named field began in the mid-1950s, the conceptual groundwork goes back further — including Alan Turing's theoretical work on computation in the 1930s and his 1950 paper proposing what became known as the Turing Test, as well as even earlier philosophical and mathematical work on formal logic and mechanical reasoning.
Do AI bootcamps actually lead to jobs?
AI bootcamps can lead to jobs, but outcomes vary widely by program and depend heavily on what the graduate does with the credential afterward — a bootcamp alone rarely gets someone hired, but it can meaningfully accelerate a transition when paired with a portfolio and targeted networking.
Do AI driven NPCs risk making games feel less predictable in a bad way?
Yes, this is a genuine, widely acknowledged risk — more open-ended AI-driven NPC behavior can produce inconsistent characterization, break narrative immersion, or generate inappropriate content in ways that feel jarring rather than immersive, which is why most developers currently pair AI-driven variation with careful guardrails rather than allowing fully unconstrained behavior.
Do AI underwriting models discriminate against protected groups?
AI underwriting models can produce discriminatory outcomes if they learn biased patterns from historical data or use data functioning as a proxy for a protected characteristic, a documented, real risk that is why insurance regulators generally require bias testing and prohibit factors producing discriminatory effects.
Do applicant tracking systems really reject resumes for formatting issues?
Yes — applicant tracking systems can genuinely fail to correctly parse resumes with complex formatting, such as tables, columns, or unusual fonts, sometimes causing relevant information to be missed even when qualifications are present, which is why experts recommend simpler formatting for online applications.
Do cloud provider AI certifications like AWS Google and Microsoft actually matter to employers?
Yes, generally — major cloud provider AI and machine learning certifications are widely recognized and often explicitly listed as preferred qualifications in job postings, particularly for roles involving deployment on that specific provider's infrastructure, though they still work best alongside demonstrated applied experience rather than as a stand-alone qualification.
Do employers care more about a certification or a portfolio project?
Most hiring signals suggest employers generally weigh a strong, relevant portfolio project more heavily than a certification alone, since a project demonstrates applied judgment and real capability, while a certification mainly demonstrates exposure to structured content — the strongest resumes typically combine both rather than relying on either alone.
Do farmers need reliable internet access for most AI agriculture tools to work?
Many AI agriculture tools do require reliable internet access, at least periodically, to transmit sensor or imagery data to cloud-based systems for analysis and to receive resulting recommendations, though some tools are designed to operate with intermittent connectivity or process data locally, making connectivity requirements an important practical consideration that varies by specific tool.
Do free AI courses lead to real job outcomes?
Yes, free AI courses can lead to real job outcomes, particularly when paired with demonstrated applied project work, since employers generally care more about actual capability than whether the underlying learning was free or paid — though free courses require more self-directed discipline to translate into a job-ready skill set.
Do police departments need a warrant to use AI facial recognition?
Warrant requirements for police use of AI facial recognition vary significantly by jurisdiction, since there's no single federal law requiring a warrant for this technology — some states and cities have enacted specific rules, including warrant requirements, while many other jurisdictions leave this legally ambiguous.
Do self-driving cars need a human safety driver by law?
Whether self-driving cars need a human safety driver by law depends significantly on the vehicle's autonomy level and jurisdiction, with lower-level systems generally requiring an attentive driver by both design and legal requirement, while some jurisdictions authorize higher-level operation without one.