The DOI matches its Crossref record; 14 of the 14 references it cites were found in a registry or an index; 6 of the 22 sources its claims cite were checked for support.
Bias in computer systems
Open accessAbstract
From an analysis of actual cases, three categories of bias in computer systems have been developed: preexisting, technical, and emergent. Preexisting bias has its roots in social institutions, practices, and attitudes. Technical bias arises from technical constraints of considerations. Emergent bias arises in a context of use. Although others have pointed to bias inparticular computer systems and have noted the general problem, we know of no comparable work that examines this phenomenon comprehensively and which offers a framework for understanding and remedying it. We conclude by suggesting that freedom from bias should by counted amoung the select set of criteria—including reliability, accuracy, and efficiency—according to which the quality of systems in use in society should be judged.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
33 Own interpretation · 6 No full text of the sources · 4 Sources not found · 3 Supported · 2 Partly supported · 2 Own method · 2 No basis given · 1 Supported so far · 1 Not checked yet · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“As specified in Section 50-(9) of the British Nationality Act itself and imple-mented in BNAP,”
§ Bias in Computer Systems > 3.3 The British Nationality Act Program (BNAP)
What the sources say
- Supportedcites Sergot et al. 1986 source
The “Formalization by Trial and Error” passage states the same distinction: a man is the father only of his legitimate children, while a woman is the mother of all her children. It also says incorporating that provision requires an adjustment to the existing rules, supporting the claim that it was implemented in the logic program. (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“At this point, it was discovered that the original Admissions Algorithm placed married couples at a disadvantage in achieving an optimal placement as compared with their single peers [Roth 1984;1990].”
What the sources say
- Partly supportedcites Roth 1984 sourceExperimental — accuracy not yet measured
Section IV, “Contemporary Issues” (page not supplied), says the couples algorithm used until 1983 could not accept preferences over pairs of positions and could produce outcomes unstable with respect to couples. That supports a disadvantage for couples, but the paper does not establish that couples achieved less optimal placements than single applicants. (LLM, conf=0.84)
- Not checked — no full text of the sourcecites Roth 1990 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“In this manner the MLSA gives processing attention to all processes as quickly as possible that are beginning a new command.”
§ Bias in Computer Systems > 3.2 A Multilevel Scheduling Algorithm (MLSA)
What the sources say
- Supportedcites Corbato et al. 1962 source
The “A Multi-Level Scheduling Algorithm” section says programs entering the system are placed in priority queues and that scheduling begins with the lowest occupied level, switching to a newly occupied lower level. The IBM 7090 system section also says the supervisor gives top priority to initiating waiting commands. Together, these passages support the paraphrase about prompt attention to processes beginning a command. (LLM, conf=0.88)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“At the time of the BNAP's initial implementation (1983) no mechanism was built into the program to incorporate relevant case law as it came into being [Sergot et al. 1986].”
§ Bias in Computer Systems > 3.3 The British Nationality Act Program (BNAP)
What the sources say
- Supportedcites Sergot et al. 1986 source
The introduction says the Act was free of case law when the original implementation began in summer 1983. It also says the authors deliberately avoided incorporating rules derived from previous cases, leaving treatment of case law for further investigation. These passages support the claim that the initial program had no built-in means of incorporating emerging case law. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The Admissions Algorithm does not take into account geographic distribution when determining a match, considering only the ranked preferences of hospitals and students.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Although others have pointed to bias in particular computer systems and have noted the general problem, we know of no comparable work that examines this phenomenon comprehensively and which offers a framework for understanding and remedying it.”
- Also stated in: § INTRODUCTION · hedged
“Although others have pointed to bias in particular computer systems and have noted the general problem [Johnson and Mulvey 1993;Moor 1985], we know of no comparable work that focuses exclusively on this phenomenon and examines it comprehensively.”
§ Bias in Computer Systems > Abstract
What the sources say
- Not matchedcites Johnson and Mulvey 1993
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not checked — no full text of the sourcecites Moor 1985 source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
- Also stated in: § INTRODUCTION · hedged
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“From an analysis of actual cases, three categories of bias in computer systems have been developed: preexisting, technical, and emergent.”
- Also stated in: § INTRODUCTION · plain
“From an analysis of actual computer systems, we have developed three categories: preexisting bias, technical bias, and emergent bias.”
- Also stated in: § 2. FRAMEWORK FOR ANALYZING BIAS IN COMPUTER SYSTEMS · plain
“The framework that emerged from this methodology is comprised of three overarching categories-preexisting bias, technical bias, and emergent bias.”
§ Bias in Computer Systems > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption: The 17 computer systems examined are representative enough to generalize the framework to computer systems generally.
- Also stated in: § INTRODUCTION · plain
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Emergent bias arises in a context of use.”
- Also stated in: § 2.3 Emergent Bias · plain
“emergent bias arises only in a context of use.”
§ Bias in Computer Systems > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 2.3 Emergent Bias · plain
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Thus, in terms of overall service from the processor, these individuals with long-running programs are systematically disadvantaged.”
§ Bias in Computer Systems > 3.2 A Multilevel Scheduling Algorithm (MLSA)
plaindirectn = 2- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Thus the British Nationality Act is biased against the illegitimate descendants of British men.”
§ Bias in Computer Systems > 3.3 The British Nationality Act Program (BNAP)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Probably not a substantial claim. 18
- It may state two separate claims as one. 15
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“For in the same text the authors also state (as noted above) that their goal is to translate the British Nationality Act so that”
§ Bias in Computer Systems > 5. CONCLUSION
What the sources say
- Partly supportedcites Sergot et al. 1986 sourceExperimental — accuracy not yet measured
Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“One charges that the Admissions Algorithm systematically favors hospital programs over medical students in cases of conflict [Graettinger and Peranson 1981b;Roth 1984;Sudarshan and Zisook 1981;Williams et al. 1981].”
§ Bias in Computer Systems > 3.1 The National Resident Match Program (NRMP)
What the sources say
- Not matchedcites Graettinger and Peranson 1981b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Roth 1984 source
Section III (Theoretical Analysis; page unavailable) states that the NRMP algorithm produces the best stable outcome for hospitals and the worst for students when preferences are submitted as strict rank-order lists. Section IV (Contemporary Issues; page unavailable) also identifies whether the algorithm unfairly favors hospitals as an issue that had been raised, while noting a qualification when participants may be indifferent between matches. (LLM, conf=0.92)
- Not checked — no full text of the sourcecites Williams et al. 1981 source
hedgeddirect- Asserted: fails — The authors may not put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The authors may not put it forward as true. The paper’s conclusions may not depend on it.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Technical bias arises from technical constraints or considerations.”
- Also stated in: § 2.1 Preexisting Bias · plain
“Technical bias arises from technical constraints or technical considerations.”
§ Bias in Computer Systems > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
- Also stated in: § 2.1 Preexisting Bias · plain
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Preexisting bias has its roots in social institutions, practices, and attitudes.”
- Also stated in: § INTRODUCTION · plain
“Preexisting bias has its roots in social institutions, practices, and attitudes.”
§ Bias in Computer Systems > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
- Also stated in: § INTRODUCTION · plain
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In this case, we ourselves would shy away from viewing this as a bias in the system because we think this discrimination can be defended as having a reasonable basis.Namely, the discrimination reflects the preferences of match participants, and it is reasonable in our view for employment decisions to be determined largely by the choices of employers and employees.”
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Rests on an assumption: The discrimination against rural hospitals reflects the preferences of match participants.
- Rests on an assumption: The best teaching hospitals tend to be in urban areas, and urban areas tend to fill their positions far more effectively, and with better students, than rural hospitals.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Emergent bias poses a potential problem for any legal expert system, especially in a society where legal systems depend on evolving case law.”
§ Bias in Computer Systems > 3.3 The British Nationality Act Program (BNAP)
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We thus infer that in this case the biased designs were unintended and, instead, reflected gender-biases deeply rooted in the culture at large.”
§ Bias in Computer Systems > 4. CONSIDERATIONS FOR MINIMIZING BIAS IN COMPUTER SYSTEM DESIGN
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As Sabre and Apollo illustrate, biases in computer systems can be difficult to identify let alone remedy because of the way the technology engages and extenuates them.Computer systems, for instance, are comparatively inexpensive to disseminate, and thus, once developed, a biased system has the potential for widespread impact.If the system becomes a standard in the field, the bias becomes pervasive”
§ Bias in Computer Systems > INTRODUCTION
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Huff and Cooper report that along a number of dimensions the designs proposed by subjects in the gender-unspecified group closely resembled the designs proposed by subjects who designed for boys and were significantly different from the designs proposed by subjects who designed for girls.”
§ Bias in Computer Systems > 4. CONSIDERATIONS FOR MINIMIZING BIAS IN COMPUTER SYSTEM DESIGN
What the sources say
- Not checked — no full text of the sourcecites Huff and Cooper 1987 source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“One source of this alleged bias lies in Sabre's and Apollo's algorithms for controlling search and display functions.In the algorithms, preference is given to”
§ Bias in Computer Systems > INTRODUCTION
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Thus, a shift in context of use may well create difficulties for a new set of users.”
§ Bias in Computer Systems > 2.2 Technical Bias
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim could be matched to a paper in Crossref, so there is nothing to check it against.
In the paper
“Thus, the computer systems systematically downgrade and, hence, are biased against international carriers who fly few, if any, internal U.S. flights, and against internal carriers who do not fly international flights [Fotos 1988;Ott 1988].”
§ Bias in Computer Systems > INTRODUCTION
What the sources say
- Not matchedcites Fotos 1988
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Ott 1988
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This bias typically emerges some time after a design is completed, as a result of changing societal knowledge, population, or cultural values.”
§ Bias in Computer Systems > 2.2 Technical Bias
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“One is that the interface design compounds the bias in the reservation systems.Lists of ranked flight options are displayed screen by screen.Each screen displays only two to five options.The advantage to a carrier of having its flights shown on the first screen is enormous since 90% of the tickets booked by travel agents are booked by the first screen display [Taib 1990].”
§ Bias in Computer Systems > INTRODUCTION
What the sources say
- Not matchedcites Taib 1990
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Or consider again the case of technical bias with the MLSA (which favors users with short jobs over long jobs).While technical bias is embedded in the program, the bias is a phenomenon of the system in use, in a context wherein users with short and long jobs outpace the system's capacity.”
§ Bias in Computer Systems > 4. CONSIDERATIONS FOR MINIMIZING BIAS IN COMPUTER SYSTEM DESIGN
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“In their study, subjects were asked to propose designs for software to teach seventh graders the correct use of commas.One group of subjects was asked to design the software for seventh-grade boys, the second group to design for seventh-grade girls, and the third group to design for seventh-graders, gender unspecified.”
§ Bias in Computer Systems > 4. CONSIDERATIONS FOR MINIMIZING BIAS IN COMPUTER SYSTEM DESIGN
What the sources say
- Not checked — no full text of the sourcecites Huff and Cooper 1987 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This type of bias can enter a system either through the explicit and conscious efforts of individuals or institutions, or implicitly and unconsciously, even in spite of the best of intentions.”
§ Bias in Computer Systems > 2.1 Preexisting Bias
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Sources of technical bias can be found in several aspects of the design process, including limitations of computer tools such as hardware, software, and peripherals; the process of ascribing social meaning to algorithms developed out of context; imperfections in pseudorandom number generation; and the attempt to make human constructs amenable to computers, when we quantify the qualitative, discret”
§ Bias in Computer Systems > 2.2 Technical Bias
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The constraint inherent in the NRMP algorithm, in which preference is given to hospital choices when conflicts in rankings occur, duplicates what happens in an actual admissions process without a computerized matching program”
§ Bias in Computer Systems > 3.1 The National Resident Match Program (NRMP)
What the sources say
- Not matchedcites Graettinger and Peranson 1981b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Whatever ranking algorithm is used, if it systematically places certain airlines' flights on initial screens and other airlines' flights on later screens, the system will exhibit technical bias.”
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“They can also reflect the personal biases of individuals who have significant input into the design of the system, such as the client or system designer.”
§ Bias in Computer Systems > 2.1 Preexisting Bias
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“those responsible for maintaining the system responded conscientiously to this societal change and modified the system's algorithm to place couples more fairly [Roth 1990].”
§ Bias in Computer Systems > 4. CONSIDERATIONS FOR MINIMIZING BIAS IN COMPUTER SYSTEM DESIGN
What the sources say
- Not checked — no full text of the sourcecites Roth 1990 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Preexisting biases may originate in society at large, in subcultures, and in formal or informal, private or public organizations and institutions.”
§ Bias in Computer Systems > 2.1 Preexisting Bias
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Should the accumulation of case law lead to changes in the way the Act is interpreted-say, by granting new subgroups British citizenship-BNAP would systematically misinform members of this subgroup regarding their status as British citizens [Leith 1986].”
§ Bias in Computer Systems > 3.3 The British Nationality Act Program (BNAP)
What the sources say
- Not checked — no full text of the sourcecites Leith 1986 source
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“In total, we examined 17 computer systems from diverse fields including banking, commerce, computer science, education, medicine, and law.”
§ Bias in Computer Systems > 2. FRAMEWORK FOR ANALYZING BIAS IN COMPUTER SYSTEMS
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“We derived our framework by examining actual computer systems for bias.Instances of bias were identified and characterized according to their source, and then the characterizations were generalized to more abstract categories.These categories were further refined by their application to other instances of bias in the same or additional computer systems.”
§ Bias in Computer Systems > 2. FRAMEWORK FOR ANALYZING BIAS IN COMPUTER SYSTEMS
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“although we have talked about bias in computer systems, the presence of bias is not so much a feature inherent in the system independent of the context of use, but an aspect of a system in use.”
§ Bias in Computer Systems > 4. CONSIDERATIONS FOR MINIMIZING BIAS IN COMPUTER SYSTEM DESIGN
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence.
The paper relies on this without support.
In cases of conflict, there is no good rationale for favoring hospital programs at the expense of medical students.
“We are inclined to answer yes because in cases of conflict there does not appear to be a good rationale for favoring hospital programs at the expense of students.”
The paper relies on this without support.
The discrimination against rural hospitals reflects the preferences of match participants.
“Namely, the discrimination reflects the preferences of match participants, and it is reasonable in our view for employment decisions to be determined largely by the choices of employers and employees.”
The paper relies on this without support.
The 17 computer systems examined are representative enough to generalize the framework to computer systems generally.
“In total, we examined 17 computer systems from diverse fields including banking, commerce, computer science, education, medicine, and law.”
The paper relies on this without support.
The illustrative examples in Table I portray plausible cases of bias.
“The illustrative examples portray plausible cases of bias.”
The paper relies on this without support.
The best teaching hospitals tend to be in urban areas, and urban areas tend to fill their positions far more effectively, and with better students, than rural hospitals.
“Because the best teaching hospitals tend to be in urban areas, the urban areas tend to fill their positions far more effectively, and with better students, than rural hospitals.”
The authors state this as an assumption.
The MLSA system is not saturated with too many users.
“assuming the system is not saturated with too many users”
The paper relies on this without support.
BNAP was designed in a research environment among people with sophisticated knowledge of immigration law.
“First, the system was designed in a research environment, among people with sophisticated knowledge of immigration law.”
The paper relies on this without support.
BNAP's users are likely to be at best paralegal or immigration counselors in Britain, if not lay persons in foreign countries with only limited access to British legal expertise.
“Its users, however, are likely to be at best paralegal or immigration counselors in Britain, if not lay persons in foreign countries with only limited access to British legal expertise.”
The paper relies on this without support.
Some of BNAP's queries require expert knowledge to answer.
“Some of the program's queries, for example, require expert knowledge to answer.”
The paper relies on this without support.
Nonexperts advising British citizen hopefuls are not alerted to alternative legal frameworks that complement the British Nationality Act.
“More generally, nonexperts advising British citizen hopefuls are not alerted to alternative legal frameworks that complement the British Nationality Act.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The multi-metric approach (Figure S1) makes explicit potential trade-offs and helps ensure the nonaccuracy desiderata are not treated as second-class citizens to accuracy (see Ref. 13).”
From Holistic Evaluation of Language Models (2023) · cites this paper as [13]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“research has repeatedly shown that these systems can in fact encode biases into decisions 124 .”
From Administrative law and the machines of government: judicial review of automated public-sector decision-making (2019) · cites this paper as B Friedman and H Nissenbaum 'Bias in Computer Systems' (1996)
The cited source says what the claim attributes to it.
In the citing paper
“The risk of”
From Positioning Paradata: A Conceptual Frame for AI Processual Documentation in Archives and Recordkeeping Contexts (2023) · cites this paper as [25]
The cited source says what the claim attributes to it.
In the citing paper
“Bias or opacity are far from newly observed characteristics of computer systems generally [26, 11, 24], but these issues have been exacerbated by the rapidly expanding numbers of data-driven information systems entering decision-making processes.”
From Fairness and Accountability Design Needs for Algorithmic Support in High-Stakes Public Sector Decision-Making (2018) · cites this paper as [26]
In the citing paper
“These decisions may be based on factors of ethical importance (e.g., ethnic, gender, or religious grounds) and irrelevant to the decision-making at hand, or they may be relevant but legally protected as a nondiscriminatory characteristic (Friedman and Nissenbaum 1996).”
From How to Design AI for Social Good: Seven Essential Factors (2020) · cites this paper as Friedman and Nissenbaum 1996
In the citing paper
“On this account, the apparent problem to which data science is a solution determines whether it happens to serve morally defensible ends.”
From Problem Formulation and Fairness (2019) · cites this paper as [17]
In the citing paper
“Corresponding with [14] this outcome systematically disadvantages individuals of one gender by preferring the other at the topk positions.”
From FA*IR: A Fair Top-k Ranking Algorithm (2017) · cites this paper as [14]
In the citing paper
“While technical factors, such as biased data and design choices, play an important role, biases can also arise from the context of use, for example, how human users interpret system outputs or from a mismatch between the capabilities and values assumed in the design of the system and those of its actual users (Danks & London, 2017; Friedman & Nissenbaum, 1996).”
From Digital ageism: challenges and opportunities in artificial intelligence for older adults (2022) · cites this paper as Friedman & Nissenbaum, 1996
In the citing paper
“and to data protection, privacy, and other human rights issues 11”
From Administrative Law and the Machines of Government: Judicial Review of Automated Public-Sector Decision-Making (2018) · cites this paper as [11]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Bias in Computer Systems > 3.1 The National Resident Match Program (NRMP)
“Table II.The Simplest Case of Rank-Order Lists in which the Desires of Students and Programs are in Conflict (Reprinted from Williams et al. [1991], by permission of the New England Journal of Medicine)”
§ Bias in Computer Systems > References
“18. WILLIAMS, K. J., WERTH, V. P., AND WOLFF, J. A. 1981. An analysis of the resident match.”
Citations
Cites
The evolution of the labor market for medical interns and residents: A case study in game theory
Roth, A. E. (1984) · Journal of Political Economy · doi:10.1086/261272 · Cited by 1,219
MOOR, J. H. (1985) · Metaphilosophy · doi:10.1111/j.1467-9973.1985.tb00173.x · Cited by 863
The British Nationality Act as a logic program
Sergot, M. J., Sadri, F., Kowalski, R. A., Kriwaczek, F., Hammond, P., & Cory, H. T. (1986) · Communications of the ACM · doi:10.1145/5689.5920 · Cited by 564
An experimental time-sharing system
Corbató, F. J., Merwin-Daggett, M., & Daley, R. C. (1962) · Proceedings of the May 1-3, 1962, spring joint computer conference on - AIEE-IRE '62 (Spring) · doi:10.1145/1460833.1460871 · Cited by 234
New physicians: A natural experiment in market organization
Roth, A. E. (1990) · Science · doi:10.1126/science.2274783 · Cited by 141
Sex Bias in Educational Software: The Effect of Designers' Stereotypes on the Software They Design1
Huff, C., & Cooper, J. (1987) · Journal of Applied Social Psychology · doi:10.1111/j.1559-1816.1987.tb00328.x · Cited by 85
Fundamental Errors in Legal Logic Programming
Leith, P. (1986) · The Computer Journal · doi:10.1093/comjnl/29.6.545 · Cited by 51
Moor, J. H. (2017) · Routledge eBooks · doi:10.4324/9781315259697-7 · Cited by 40
Moor, J. H. (2020) · Routledge eBooks · doi:10.4324/9781003075011-1 · Cited by 21
National resident matching program
(1981) · New England Journal of Medicine · doi:10.1056/nejm198108273050923 · Cited by 10
Research works printed but not linked: 4
- FISHLOCK , T. 1981. Delhi press detect racism in Nationality Bill. The London Times ( Jan. 20 ) . FISHLOCK, T. 1981. Delhi press detect racism in Nationality Bill. The London Times (Jan. 20).
- GAO . 1992. Patriot Missile defense: Software problem led to system failure at Dhahran , Saudi Arabia. GAO/IMTEC-92-26, U.S. General Accounting Office, Washington, D.C. GAO. 1992. Patriot Missile defense: Software problem led to system failure at Dhahran, Saudi Arabia. GAO/IMTEC-92-26, U.S. General Accounting Office, Washington, D.C.
- JOHNSON , D.G. AND MULVEY , J. M. 1993. Computer decisions: Ethical issues of responsibil-ity and bias. Statistics and Operations Res. Series SOR-93-11 , Dept. of Civil Engineering and Operations Research , Princeton Univ ., Princeton, N.J. JOHNSON,D.G.AND MULVEY, J. M. 1993. Computer decisions: Ethical issues of responsibil-ity and bias. Statistics and Operations Res. Series SOR-93-11, Dept. of Civil Engineering and Operations Research, Princeton Univ., Princeton, N.J.
- OTT , J. 1988. American Airlines settles CRS dispute with British Airways. Aviat. Week Space Tech. ( July 18 ) . OTT, J. 1988. American Airlines settles CRS dispute with British Airways. Aviat. Week Space Tech. (July 18).
Cited by
Sorted by citations
Items 1,101–1,155 of 1,155
Rubeis, G. (2026) · Springer eBooks · doi:10.1007/978-3-032-14301-3_5
Hegberg, A., Setor, T. K., & Joseph, D. (2026) · ACM SIGCPR Conference on Computer Personnel Research (SIGCPR) · doi:10.1145/3768310.3807790 · Cited by 1
“It can show up as absence or as templated presence, consistent with lifecycle accounts showing that harms often emerge upstream of observable allocations (Chien & Danks, 2024; Friedman & Nissenbaum, 1996; Suresh & Guttag, 2021).”
InfluentialRegulierung, Governance und Technikrecht
Keuthen, T. (2026) · Springer eBooks · doi:10.1007/978-3-658-51418-1_9
Human-Centered and Participatory AI Auditing
Deng, W. H., Holstein, K., & Eslami, M. (2026) · Springer eBooks · doi:10.1007/978-981-95-3658-0_102
Enabling Diversity and Gender Equity in Human-Centered AI
Anderson, T. D., & Marshall, R. (2026) · Springer eBooks · doi:10.1007/978-981-95-3658-0_104
121Chapter 3 Why AI requires better regulation: Ten compelling arguments
(2026) · De Gruyter eBooks · doi:10.1515/9783112245958-004
Sources of Bias in Clinical Artificial Intelligence and Applications in Rheumatology
Creasman, M., García-Agúndez, A., Yazdany, J., & Schmajuk, G. (2026) · Rheumatic Disease Clinics of North America · doi:10.1016/j.rdc.2026.03.005
Matzner, T. (2026) · Springer eBooks · doi:10.1007/978-3-658-36625-4_5
Rohrbach, T., Makhortykh, M., & Sydorova, M. (2026) · Political Communication · doi:10.1080/10584609.2026.2679476
COIN Enterprise Transformation
Ramaswamy, V., Ozcan, K., & Narayanan, K. (2026) · Springer eBooks · doi:10.1007/978-3-032-19468-8_6
Regulating the Emergence of AI in Education
More, A. B., & Shrivastava, A. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7467-3.ch005
Bunde, E. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-28316-0_12
Detecting What AI Metrics Miss: A Diagnostic Test for Value-Level Failures
Combe, C. G. (2026) · IEEE Swiss Conference on Data Science (SDS) · doi:10.1109/sds70563.2026.00010
Agarwal, S., Singh, H. K., Sah, M. K., et al. (2025) · International Conference on Information Management & Machine Intelligence (ICIMMI) · doi:10.1145/3793449.3793479
“Third, feedback loops turn small biases into systemic unfairness by continuously reinforcing what the system already believes [10-11].”
Meyer, G., Schroeder, F., & Rainer, A. (2026) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-032-26999-7_6
Ayan, M. A. (2026) · Gazi Üniversitesi Fen Bilimleri Dergisi Part C Tasarım ve Teknoloji · doi:10.29109/gujsc.1924721
The Moral-Constraint Divide in AI Development
Bailey, P. (2026) · IGI Global eBooks · doi:10.4018/979-8-3373-8192-3.ch010
A New Perspective on AI Ethics Education for a Future of Human-AI Teammates
Strug, B., Ślusarczyk, G., & Strug, J. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-29791-4_37
Stinar, F., Yao, C., Marras, M., Yu, R., & Bosch, N. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-29794-5_15
Fahim, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7011188
Havens, L., Terras, M., Bach, B., & Alex, B. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812283
“Friedman and Nissenbaum [53] categorize these types of computational biases as “preexisting bias,” because the biases existed in institutional and personal practices and attitudes prior to the technology’s creation.”
VISUAL PERCEPTION IN POST‐SCREEN CULTURES IN MIXED REALITY INTERFACESPreprint
Fahim, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7038321
Ellis, A. L., & Bryant, J. (2026) · Educational Policy · doi:10.1177/08959048261465623
Cao, S., Chen, Z., & Wang, H. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2700023
Tyulina, N. (2026) · ACM Conference on Conversational User Interfaces (CUI) · doi:10.1145/3816046.3816313
“Previous work has documented epistemic harms [5], racial bias [22], systematic bias [14], bias toward disabled communities [15], risks of psychological manipulation [1], plagiarism and replication of stereotypes [13], and clear evidence of harm in the form of inappropriate sexual behavior [19].”
A survey of gender bias mitigation in neural machine translation
Gajakos, N., Staff, C., Murphy, B., Kelleher, J. D., & Haque, R. (2026) · Language Resources and Evaluation · doi:10.1007/s10579-026-09931-x
“According to Friedman and Nissenbaum (1996), as cited by Savoldi et al. (2021), these phenomena fit into the broader categories of pre-existing, technical, and emer - gent bias, as further discussed in Sect.”
Afisi, O. T., Olabanjo, O. A., & Wusu, A. S. (2026) · Journal of Artificial Intelligence · doi:10.3923/jai.2026.60.71
Chen, W. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7017078
Bias in AI: a framework for understanding the sources of bias in artificial intelligence
Robinig, W., & Wallner, J. P. (2026) · AI and Ethics · doi:10.1007/s43681-026-01283-0 · Cited by 1
“A model trained on such data inherits the categories and assumptions of the period from which the data was drawn, and may reproduce them in contexts where they no longer apply [ 12 , 30].”
InfluentialPerspectives of Healthcare Professionals on Using AI and Algorithmic Bias
Xu, J., & Babaian, T. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-34459-5_1
Can AI reduce gender bias? The case of AI-generated audio conversations in Google's NotebookLM
Malekzadeh, M. (2026) · Computers in Human Behavior Reports · doi:10.1016/j.chbr.2026.101246
Walkiewicz, J. (2026) · Wroclaw University of Economics and Business eBooks · doi:10.15611/2026.53.1.04
Fairness, Bias, and Equity in Algorithmic Systems
Cook, D. (2026) · Springer eBooks · doi:10.1007/978-3-031-52643-5_263-1
PROPOSTA DE MODELO PARA DIMINUIR O VIÉS ALGORÍTMICO NA SELEÇÃO DE PESSOAS APOIADAS POR IA
de Andrade, R. A., Cassarotti, L. C., Sewald, E., Corrêa, U. C., Velloso, B. P., Cardozo, G., & Bueno, L. (2026) · Revista de Geopolítica · doi:10.56238/revgeov17n8-147 · Cited by 1
Yapay Zekâ ve İnsan Zekâsı İş Birliğiyle Kamu Değeri Yaratımı: Hibrit Yönetişim Modellerinin Rolü
Uzun, H. (2026) · Ankara Hacı Bayram Veli Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi · doi:10.26745/ahbvuibfd.1763217
Algorithmic fairness: theorems of incompleteness and the limits of formalizing normative concepts
Chekrygin, A. D. (2026) · Философская мысль · doi:10.25136/2409-8728.2026.8.81840
Mühlhoff, R. (2026) · AI & Society · doi:10.1007/s00146-026-03281-6
Poechhacker, N., & Klausner, L. D. (2026) · De Gruyter eBooks · doi:10.1515/9783111373096-026
Davidian, M., Mazliah, D., Slovin, G., Joshua, B., Lurie, Y., & Mark, S. (2026) · Software · doi:10.3390/software5030039
“Dignum [28] offers a typology grounded in the insight that technology is not value-neutral but actively embodies values [15,16].”
ALGORİTMİK ÖN YARGININ DİJİTAL HAKLAR ÜZERİNDEKİ ETKİLERİ
Çelik, D. B. (2026) · Karabük Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi · doi:10.63233/kbuiibf.1988678
When algorithms discriminate: AI bias and its implications for tourism and hospitality
Solarević, M., Kalinić, Č., Vujičić, M. D., & Stankov, U. (2026) · Zbornik radova Departmana za geografiju turizam i hotelijerstvo · doi:10.5937/zbdght2601096s
“One of the key areas of risk is bias in AI, defined as unfair discrimination against individuals or groups at the expense of others (Friedman & Nissenbaum, 1996).”
Visual Representation of "Chinoiserie" in Western AI Image Models: A Case Study of Midjourney
Wang, S. (2026) · International Journal of Social Science Education and Humanities · doi:10.63313/ijsseh.9063
Responsible AI in healthcare: Navigating ethics and governance challenges
Vithya, N., Subha, N., Nisha, S. P., & Balthilak, A. (2026) · Elsevier eBooks · doi:10.1016/b978-0-443-27761-0.00005-6
Berendsen, S., & Beauxis-Aussalet, E. (2026) · AI and Ethics · doi:10.1007/s43681-026-01375-x
Feedback-driven rule induction and retrieval-augmented bias mitigation for large language models
Kumar, A., Shree, M. A., Balaji, V. K., & Bhargavi, R. C. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1925701
“Although LLMs achieve strong performance across generation and reasoning tasks, they often reproduce societal biases present in training data, including gendered associations and stereotypical reasoning patterns (Friedman and Nissenbaum, 1996; Caliskan et al., 2017; Wan et al., 2023).”
Gallardo, A., Klucinec, L., Bauer, L., & Cranor, L. F. (2026) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3817022
“Previous work on harms of algorithmic systems conveys how marginalized communities disproportionately experience sociotechnical harms [38, 41].”
Which AI? A Comparative Evaluation of Generative AI and Neuro‐Symbolic AI for Silicon Sampling
Zahradníčková, K., Polišenský, J., Lukeš, M., Hranický, R., & Pouč, P. (2026) · Psychology and Marketing · doi:10.1002/mar.70272
Han, C., Wang, S.-M., Wang, C.-C., Sun, J.-S., & Fang, T.-J. (2026) · Frontiers in Nutrition · doi:10.3389/fnut.2026.1799873
Safety Considerations During Rapid Implementation of A.I. in India
Supriya, B. (2026) · Indian Journal of Social Science and Literature · doi:10.54105/ijssl.a1246.06010926
RetractedCascaes, F. E. D. M. (2026) · Revista Multidisciplinar do Nordeste Mineiro · doi:10.66104/aczpy604
Detecting and Mitigating Bias in AI Systems: A Methodological Approach to Social Inequality
Grimaccia, E. (2026) · Societies · doi:10.3390/soc16100316
YAPAY ZEKA VE ALGORİTMİK YANLILIK: TOPLUMSAL HABERLER ÜZERİNDEN BİR İNCELEME
ÖNAL, A., & DİLMEN, N. E. (2026) · Gümüşhane Üniversitesi İletişim Fakültesi Elektronik Dergisi · doi:10.19145/e-gifder.1856322
Liu, J., Zhong, C., Meng, F., Shank, D. B., & Wunsch, D. C. (2026) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2026.103140
Metadata
- Authors
- Friedman, B., & Nissenbaum, H.
- Affiliations
- Colby College (US)
- Princeton University (US)
- Venue
- ACM Transactions on Information Systems
- Published
- 1996-07
- Volume
- 14
- Issue
- 3
- Pages
- 330-347
- ISSN
- 1046-8188 (ISSN-L)
- 1558-2868
- OpenAlex
- W2006447892
- Cited by
- 1,203
- References
- 13
- Citations by year
- 2026: 102 · 2025: 149 · 2024: 151 · 2023: 121 · 2022: 126 · 2021: 133 · 2020: 79 · 2019: 70 · 2018: 50 · 2017: 21 · 2016: 29 · 2015: 18 · 2014: 18 · 2013: 16 · 2012: 8
- FWCI
- 17.25
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- ACM Transactions on Information Systems (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Social Sciences › Business, Management and Accounting › Management Information Systems › Information Systems Theories and Implementation
- Social Sciences › Law › Artificial Intelligence in Law
- Keywords
- technical bias, social institutions, technological constraints, societal impact of technology
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
