2,554 published verifications avg. score 5.6/10 1,246 rated true or mostly true 1,065 rated false or mostly false
“Substantive disagreements between AI models on fact-checking outcomes are common.”
Evidence from multiple studies shows that AI fact-checking models often reach materially different verdicts on the same claim, with reported substantive conflicts commonly in the roughly 15% to 30% range on challenging datasets. That is frequent enough to count as common in real-world use. Rates do vary by claim difficulty, ambiguity, prompting, and evidence quality.
“Capitalism tends to promote the natural emergence of monopolies.”
There is solid evidence that many capitalist economies have seen rising concentration and market power, especially in recent decades. But that does not establish that capitalism inherently or naturally produces monopolies. The claim overstates the evidence by treating concentration as monopoly and by omitting the major roles of technology, policy, and antitrust enforcement.
“Portugal's digital nomad visa provides a direct pathway to Portuguese citizenship.”
Portugal’s digital nomad regime does not create a special citizenship track. The residence-permit version can count toward the standard five-year residence requirement for naturalization, but citizenship is a separate, conditional process under nationality law. The claim is misleading because it suggests a built-in or streamlined route and ignores that the temporary-stay version does not lead to citizenship at all.
“Using prescription acne medication at any point during pregnancy is problematic for pregnant women.”
The evidence does not support a blanket claim that prescription acne medication is problematic at any point in pregnancy. Major medical guidance says risk is drug-specific: retinoids are contraindicated, while several prescription options, such as topical clindamycin and azelaic acid, are commonly considered acceptable in pregnancy. The statement wrongly treats the entire category as unsafe.
“Alternative medicine works as well as or better than conventional medicine.”
The evidence does not support the idea that alternative medicine broadly works as well as or better than conventional medicine. A few approaches, such as acupuncture for some pain conditions, may offer modest benefit, but they generally do not outperform standard care. Many others, especially homeopathy, fail rigorous testing, and replacing proven treatment with alternatives can increase harm and mortality.
“Sugar-free soft drinks sweetened with non-sugar sweeteners (for example, Coca-Cola Zero Sugar) are as harmful to teeth as sugar-sweetened soft drinks.”
The evidence does not support a blanket claim of equal dental harm. Sugar-free soft drinks can damage teeth through acid-related enamel erosion at levels that may be similar to sugary soft drinks, but sugar-sweetened drinks also raise cavity risk through fermentable sugars. That extra cariogenic pathway makes sugary drinks generally more harmful overall to teeth.
“Salivary amylase cannot break down starch into glucose because the enzyme molecule is too large.”
The evidence shows salivary amylase does break down starch, just not primarily into free glucose. It begins starch digestion in the mouth by producing smaller sugars such as maltose and dextrins. The stated reason is also wrong: no credible source supports the idea that amylase fails to make glucose because the enzyme molecule is “too large.”
“A height of 6 feet 2 inches is approximately the 95th percentile for adult male height in the United States.”
The claim is well supported by U.S. anthropometric data. CDC/NCHS NHANES references place the 95th percentile for adult U.S. men at about 187.3-187.5 cm, while some related analyses put it near 188 cm. Since 6'2" equals 188.0 cm, describing it as approximately the 95th percentile is accurate.
“Win probability models used in sports analytics (for example, ESPN's NBA win probability model) are generally well-calibrated, and their probabilities align reasonably well with observed outcomes, at least compared with typical human intuition.”
The available evidence indicates these win-probability models are broadly calibrated over many games and usually align reasonably well with actual outcomes. They are not perfect: some studies find underdog bias and phase-specific quirks, and they are not clearly superior to simple baseline models. The comparison to human intuition is supported more indirectly than directly, but the overall claim is largely accurate.
“The 2020 United States presidential election was stolen.”
The evidence does not support that the 2020 U.S. presidential election was stolen. Courts, election officials, and federal agencies repeatedly found no credible evidence of widespread fraud or manipulation that could have changed the outcome, and the results were lawfully certified. Isolated irregularities and investigations did not amount to a national scheme or an illegitimate result.
“Poland is a sovereign state with full legal authority over its domestic governance and foreign policy.”
Poland is a sovereign state that governs itself under its own constitution and conducts its own foreign relations. Constitutional texts, official state practice, and international recognition strongly support that core point. However, "full legal authority" overstates the position because Poland has accepted binding limits through EU membership and other international obligations.
“In medieval Europe, there were roughly 80 to 100 feast days per year on which work generally stopped.”
The core point holds: medieval Europe had many religious holidays on which ordinary work was often restricted. But 80–100 is not a fixed Europe-wide number; it is a rough estimate that varies by region, century, and counting method, with credible figures both below and above that range. Treat it as an approximation, not a settled annual total.
“Fever can sometimes cure cancer.”
Rare cancer regressions have been reported after high fevers, acute infections, or fever-inducing treatments, so the idea has some factual basis. But the evidence mostly consists of case reports, historical series, and correlations, not proof that fever itself directly cures cancer. In modern medicine, hyperthermia is used as an adjunct to other treatments, not as a reliable stand-alone cure.
“In humans, when not eating, the entire esophagus is completely closed along its full length, so water cannot drip down into the stomach on its own.”
The claim overstates normal esophageal physiology. At rest, the esophagus is usually collapsed and its sphincters are generally closed, which is why swallowing is normally needed. But it is not a completely sealed tube along its entire length, and spontaneous lower sphincter relaxations and small amounts of intraluminal air mean the absolute wording is medically inaccurate.
“Chariots have been found at the bottom of the Red Sea, proving that the Israelites crossed it after the waters parted.”
No verified archaeological evidence shows that chariot wheels were found on the Red Sea floor. The claim relies on unverified Ron Wyatt-style reports, ambiguous underwater images, and misidentified coral formations, not recovered and authenticated artifacts. Without independent analysis or peer-reviewed documentation, it does not prove an Israelite crossing or parted waters.
“Taking 5,000 IU of vitamin D daily does not cause side effects for most people.”
Most adults are unlikely to develop side effects from 5,000 IU of vitamin D daily, and several studies found it was generally well tolerated. But the claim overstates the evidence: 5,000 IU is above the usual adult upper limit of 4,000 IU, and review-level evidence shows a small but real increase in some adverse events at nearby doses. A safer wording would be that side effects are uncommon, not absent.
“A meta-analysis published in 2023 found that taking 3,200 to 4,000 International Units (IU) of vitamin D daily is associated with a higher risk of hypercalcemia.”
The cited 2023 meta-analysis did report that long-term vitamin D supplementation at 3,200-4,000 IU/day was associated with a higher risk of hypercalcemia. The pooled estimate showed a statistically significant increase versus control. Important context is that the absolute excess risk was small and reported cases were often asymptomatic or reversible.
“In the United Kingdom, a company's ultimate beneficial owner (UBO) can be obscured by using nominee shareholders.”
Nominee shareholders can obscure who appears as the legal shareholder of a UK company, and UK authorities acknowledge they may be used to create distance from the real owner. But that is not the full legal picture: the PSC regime is designed to look through nominees and require disclosure of the real controller when control thresholds are met. So the claim is only accurate in a limited, partial sense.
“Scientists in relevant fields widely agree that the health and environmental effects of the 1979 Three Mile Island nuclear accident were very low.”
The evidence strongly supports a broad scientific view that Three Mile Island caused very low off-site radiological health and environmental harm. Government reviews, academy assessments, and long-term epidemiology generally found exposures near background and no detectable population-level physical effects. A few studies raised limited cancer-signal concerns, and the claim is broad enough to understate psychosocial impacts, so some caution is warranted.
“Islam was forbidden in France during the French Third Republic (1870–1940).”
No evidence shows that the French Third Republic legally forbade Islam. The historical record instead shows legal protection for religious exercise in principle, alongside heavy state control and unequal treatment of Muslims, especially in colonial territories. Those restrictions were real, but they were not the same as banning the religion.