Data governance: Organizing data for trustworthy Artificial Intelligence

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Data quality consists of many dimensions including accuracy, timeliness, completeness, consistency, objectivity, believability and relevance, which all determine whether data is fit for use.

“Data quality consists of many dimensions including accuracy, timeliness, completeness, consistency, objectivity, believability and relevance, which all determine whether data is fit for use (Strong, Lee, & Wang, 1997).”

• cites Strong, Lee, & Wang, 1997 sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The choice of the data governance approach is critical, albeit this choice is not always clear or even explicit.

“Given such challenges, the choice of the data governance approach is critical, albeit this choice is not always clear or even explicit (Koltay, 2016).”

• cites Koltay, 2016 sourcePartly supported

  In “Data governance in detail” (page unavailable), Koltay says data governance should not be optional and that cross-functional perspectives on it vary. This supports its importance and suggests some ambiguity, but the passages do not say that choosing a governance approach is not always clear or explicit. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

The risk-based approach is often advocated as a foundation for data governance.

“This approach is often advocated as a foundation for data governance (Rothstein, Borraz, & Huber, 2013)”

• cites Rothstein, Borraz, & Huber, 2013 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The organizational approach to data governance emphasizes structure, responsibility, accountability and reporting.

“The organizational approach to data governance emphasizes structure, responsibility, accountability and reporting (Mullon & Ngoepe, 2019).”

• cites Mullon & Ngoepe, 2019 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S4 (based).

The mismatch between organizational structure and data usage can easily result in data silos, duplications, unclear responsibilities, and missing control of data over its entire life-cycle.

“The mismatch between organizational structure and data usage can easily result in data silos, duplications, unclear responsibilities, and missing control of data over its entire life-cycle.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Too much governance can result in excessive overhead and lower performance.

“too much governance can result in excessive overhead and lower performance.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Ad-hoc handling of data without procedures and processes and secure data infrastructure might easily result in individual data items being accessible to non-authorized persons.

“Another is ad-hoc handling of data without procedures and processes and secure data infrastructure, which might easily result in individual data items being accessible to non-authorized persons.”

Classified as probably not a substantial claim.

Beyond a single organization, data governance depends on collaboration between organizations and persons that make up the system.

“Beyond a single organization, data governance depends on collaboration between organizations and persons that make up the system.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Risk-assessment, defining responsibilities, and joint planning and control can be done between organizations, but require the establishment of trusted frameworks.

“Risk-assement, defining responsibilities, and joint planning and control can be done between organizations, but require the establishment of trusted frameworks.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The planning and control approach to data governance is often criticized for not easily adapting to change.

“The approach is often criticized for not easily adapting to change (Janssen & van der Voort, 2016).”

• cites Janssen & van der Voort, 2016 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Too little governance might result in unclear responsibilities, uncontrolled risks and not taking the right precautions and actions.

“Too little governance, however, might result in unclear responsibilities, uncontrolled risks and not taking the right precautions and actions.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Another risk area is poor generalization of the algorithms to situations outside the data upon which they were trained.

“Another risk area is poor generalization of the algorithms to situations outside the data upon which they were trained.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The risk-based approach has gained attention as a way to identify risks of BDAS and introduce appropriate governance mechanisms to address them.

“With the advent of GDPR and AI, the risk-based approach has gained attention as a way to identify risks of BDAS and introduce appropriate governance mechanisms to address them (Ladley, 2019).”

• cites Ladley, 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S4 (based).

The risks of training algorithms using historical data are that mistakes, inconsistencies and bias embedded in such data will be reflected in the working of the algorithms.

“The risks of training algorithms using historical data are that mistakes, inconsistencies and bias embedded in such data will be reflected in the working of the algorithms.”

Probably not a substantial claim: may not pass S4 (based).

Data that is not collected cannot be misused or abused.

“As data that is not collected cannot be misused or abused, the collection of data should be minimized.”

Probably not a substantial claim: may not pass S4 (based).

Low data quality and unclear dependencies between data and algorithms can easily bias or skew the outcomes of the AI algorithms.

“Low data quality and unclear dependencies between data and algorithms can easily bias or skew the outcomes of the AI algorithms.”

Probably not a substantial claim: may not pass S4 (based).

Shared roles and joint operations performed among departments and organizations may dilute responsibilities.

“Shared roles and joint operations performed among departments and organizations may dilute responsibilities.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The usefulness principle could result in data protection being violated by data stewards.

“The other way around, the usefulness principes could result in data protection being violated by data stewards.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A common challenge with data governance is that the data flow and logic may not follow the structure of an organization.

“A common challenge with data governance is that the data flow and logic may not follow the structure of an organization.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Data governance contains mechanisms to incentivize correct behaviours and sanction incorrect ones.

“Data governance contains mechanisms to incentivize correct behaviours and sanction incorrect ones.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Opening algorithms might reveal bugs in the working systems, which could be exploited by hackers.

“However, the opening of algorithms should be done with care as it might reveal bugs in the working systems, which could be exploited by hackers.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Expectations and values are translated in the planning and control, organizational, or risk-based governance approaches, resulting in the assignment of decision-making authority and the establishment of processes and procedures.

“These expectations and values are translated in the planning and control, organizational, or risk-based governace approaches, resulting in the assignment of decision-making authority and the establishment of processes and procedures.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Mechanisms are needed to ensure responsible data sharing, allowing data sharing when required, but also blocking data sharing when necessary.

“However, mechanisms are needed to ensure responsible data sharing, allowing data sharing when required, but also blocking data sharing when necessary.”

Probably not a substantial claim: may not pass S4 (based).

If sensitive data, like gender, race, residential address, health status or political preference is collected, then it can be misused or abused.

“If sensitive data, like gender, race, residential address, health status or political preference is collected, then it can be misused or abused.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Dawes (2010) formulated the stewardship and usefulness principles for information sharing: stewardship should assure data quality, validity and security; manage risks; manage and preserve data; and make public officials and organizations responsible for handling information with care and integrity.

“Dawes (2010) formulated the stewardship and usefulness principles for information sharing: stewardship should assure data quality, validity and security; manage risks; manage and preserve data; and make public officials and organizations responsible for handling information with care and integrity.”

• cites Dawes (2010) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

AI capabilities generate immense powers, which are dangerous if systems are allowed to expand without proper oversight, accountability and governance.

“AI capabilities generate immense powers, which are dangerous if systems are allowed to expand without proper oversight, accountability and governance.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Plotkin (2013) views data stewardship as a way to formalize accountabilities for managing information resources on behalf of and in the best interest of others.

“Plotkin (2013) view data stewardship as a way to formalize accountabilities for managing information resources on behalf of and in the best interest of others.”

• cites Plotkin (2013)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S2 (checkable).

Trusted data sharing frameworks are required for this.

“For this, trusted data sharing frameworks are required.”

Classified as probably not a substantial claim.

Applying BDAS in consequential decision-making is conditional on data quality.

“While applying BDAS in consequential decision-making is conditional on data quality, perfect data quality does not exist.”

• AssumptionRests on an assumption: Perfect data quality does not exist.

Classified as probably not a substantial claim.

Opening AI algorithms to controlled groups allows auditors, experts, scientists, citizen scientists and other professionals to check the working of the algorithms.

“While most people will be unable to make much sense of the algorithms, this would allow auditors, experts, scientists, citizen scientists and other professionals, equipped with proper tools, to check the working of the algorithms.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Three approaches to data governance are planning and control, organizational, and risk-based.

“Three approaches to data governance are planning and control, organizational, and risk-based.”

Probably not a substantial claim: may not pass S2 (checkable).

Even if data is pooled and linked at the conceptual level, physical storage and responsibility for data should be distributed to reduce vulnerability.

“Even if data is pooled and linked at the conceptual level, physical storage and responsibility for data should be distributed to reduce vulnerability.”

Classified as probably not a substantial claim.

The causal relationship between an event and a system failure might be difficult to establish without proper data governance.

“The causal relationship between an event and a system failure might be difficult to establish without proper data governance.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Controls can play similar preventive and detective roles, guided by socio-technical arrangements within the organizations concerned.

“Controls can play similar preventive and detective roles, guided by socio-technical arrangements within the organizations concerned.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

There might be a lack of established mechanisms for data governance for dealing with BDAS while involving various organizations.

“There might be a lack of established mechanisms for data governance for dealing with BDAS while involving various organizations.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

Creating sound data governance requires a balance between complete control, which is unreachable, and lack of control, which is dangerous, into the mode of governance that is necessary and feasible.

“Creating sound data governance requires a balance between complete control, which is unreachable, and lack of control, which is dangerous, into the mode of governance that is necessary and feasible.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

There are numerous situations where sensitive data must be collected and shared for the sake of transparency, fraud detection, service improvements, or better decisions.

“Nevertheless, there are numerous situations where sensitive data must be collected and shared for the sake of transparency, fraud detection, service improvements, or better decisions.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The more controls are embedded in governance processes, the higher auditability and problem detection rates, and the more opportunities for learning and improvement.

“The more controls are embedded in governance processes, the higher auditability and problem detection rates, and the more opportunities for learning and improvement.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

In the case of data, the notion of stewardship is preferred over ownership.

“In the case of data, the notion of stewardship is preferred over ownership.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Dawes (2010) further argued that stewardship and usefulness are complementary principles and should guide information-based transparency.

“Dawes (2010) further argued that stewardship and usefulness are complementary principles and should guide information-based transparency.”

• cites Dawes (2010) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Data governance is based on the expectations and values of the society, which is ultimately affected by the outcomes of BDAS.

“Data governance is based on the expectations and values of the society, which is ultimately affected by the outcomes of BDAS.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

The purpose of data governance is to increase the value of data and minimize data-related costs and risks.

“The purpose is to increase the value of data and minimize data-related costs and risks (Abraham, Schneider, & vom Brocke, J., 2019).”

• cites Abraham, Schneider, & vom Brocke, J., 2019 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In law, tangible goods are more natural to own than intangible ones, like data which can be shared without limits.

“Furthermore, in law, tangible goods are more natural to own than intangible ones, like data which can be shared without limits.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Stewards are also responsible for information security.

“Others emphasize that stewards are also responsible for information security (Cuganesan, Hart, & Steele, 2017).”

• cites Cuganesan, Hart, & Steele, 2017 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Defining and sharing responsibilities between data providers, algorithms provides and BDAS operators as part of data governance is a challenge, particularly when multiple organizations engage in such governance jointly.

“This raises the challenge of defining and sharing responsibilities between data providers, algorithms provides and BDAS operators as part of data governance, particularly when multiple organizations engage in such governance jointly.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Risk-based approaches to data governance are suited to maximize the value created from data while mitigating risks and reducing costs.

“Risk-based approaches to data governance are suited to maximize the value created from data while mitigating risks and reducing costs.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The ownership of data is often challenging to establish, and multiple persons might claim the rights to it.

“The ownership of data is often challenging to establish, and multiple persons might claim the rights to it.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

The foundation of data governance is responsible data collection.

“The foundation of data governance is responsible data collection.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Data is the basis for BDAS.

“Data is the basis for BDAS, but the outcomes of BDAS should be monitored as well, as they are part of the data life-cycle.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The three data governance approaches are not mutually exclusive and can be used to complement each other.

“These approaches are not mutually exclusive and can be used to complement each other.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Incentives cannot replace proper policing by audit organizations.

“However, incentives cannot replace proper policing by audit organizations.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Data stewardship and base registries provide a foundation for data and information sharing.

“Data stewardship and base registries provide a foundation for data and information sharing.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The stewardship principle might deal with information sharing risks by withholding that information and thus reducing transparency and accountability.

“The stewardship principle might deal with information sharing risks by withholding that information and thus reducing transparency and accountability.”

Probably not a substantial claim: may not pass S2 (checkable).

To avoid the situation of a single entity exercising control over data without others' consent, distributed responsibilities and separation of concerns make systems less vulnerable.

“To avoid the situation of a single entity exercising control over data without others' consent, distributed responsibilities and separation of concerns make systems less vulnerable.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Data stewardship is a team effort where responsibilities and expertise are divided among members, who can manage data on behalf of others.

“Data stewardship is a team effort where responsibilities and expertise are divided among members, who can manage data on behalf of others.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The second involves users exercising Self-Sovereign Identity (SSI).

“The second involves users exercising Self-Sovereign Identity (SSI).”

Probably not a substantial claim: may not pass S3 (load-bearing).

This reduces the chances of using the data for unrelated purposes.

“This reduces the chances of using the data for unrelated purposes.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Access to their own data not only results in transparency but also empowers citizens and organizations to be in control of their data, including data about themselves.

“Furthermore, access to their own data not only results in transparency but also empowers citizens and organizations to be in control of their data, including data about themselves.”

Probably not a substantial claim: may not pass S2 (checkable).

Such a decentralized information sharing system needs to be guided by proper governance mechanisms.

“Nevertheless, such a decentralized information sharing system needs to be guided by proper governance mechanisms.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

There is a need for technologies like base registries and self-sovereign identities to make trustworthy BDAS work.

“However, there is a need for technologies like base registries and self-sovereign identities to make trustworthy BDAS work.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Therefore organizations collect and store identity information such as name, gender, age, profession and others to be able to identify persons.

“Therefore organizations collect and store identity information such as name, gender, age, profession and others to be able to identify persons.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

The realization of data governance in multi-organizational settings will also require the use of trusted data sharing frameworks to guide inter-organizational data exchange, and ensure compliance with regulations as well as the creation of public value.

“The realization of data governance in multi-organizational settings will also require the use of trusted datasharing frameworks to guide inter-organizational data exchange, and ensure compliance with regulations as well as the creation of public value.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

This extension relies on trusted data sharing frameworks, or trusted frameworks for short, that guide data exchange within and between participating organizations while ensuring compliance with regulations and the realization of public value.

“This extension relies on trusted data sharing frameworks, or trusted frameworks for short, that guide data exchange within and between participating organizations while ensuring compliance with regulations and the realization of public value.”

Classified as probably not a substantial claim.

Organizations or users can control the data.

“Organizations or users can control the data.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

While the foundation of BDAS data governance is responsible data collection, citizens' control of data, and data stewardship, this new research field has to continue advancing before it creates a solid research foundation for trusted BDAS.

“While the foundation of BDAS data governance is responsible data collection, citizens' control of data, and data stewardship, this new research field has to continue advancing before it creates a solid research foundation for trusted BDAS.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

This also requires agreements to regulate access, openness and inclusion.

“However, this also requires agreements to regulate access, openness and inclusion.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Trusted frameworks are a standard mechanism for governing relationships and are used in payment systems, in domain registration systems, or in mobile networks to provide roaming services.

“Trusted frameworks are a standard mechanism for governing relationships and are used in payment systems, in domain registration systems, or in mobile networks to provide roaming services.”

Probably not a substantial claim: may not pass S2 (checkable).

For that, system-level governance is needed.

“For that, system-level governance is needed.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Available approaches to data governance can be based on clear organizational structures, responsibilities and accountabilities, planning and control cycles, and risks.

“Available approaches to data governance can be based on clear organizational structures, responsibilities and accountabilities, planning and control cycles, and risks.”

Classified as probably not a substantial claim. May combine two claims.

Trusted frameworks limit access to data to those who are authorized, and ensure non-repudiation of the data origin.

“Trusted frameworks limit access to data to those who are authorized, and ensure non-repudiation of the data origin.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Increasingly, trusted frameworks which are based on agreements among participating parties are used for data sharing and service provision.

“Increasingly, trusted frameworks which are based on agreements among participating parties are used for data sharing and service provision.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Data governance for BDAS is a complex field, and the development of BDAS without due attention to data governance is a significant risk.

“Data governance for BDAS is a complex field, and the development of BDAS without due attention to data governance is a significant risk.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

Data governance can help mitigate the issues of transparency, accountability, fairness, discrimination, and trust.

“Data governance can help mitigate the issues of transparency, accountability, fairness, discrimination, and trust.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

As DLT ensures decentralized execution, avoiding the creation of a single point of failure or misuse, they are less vulnerable.

“As DLT ensures decentralized execution, avoiding the creation of a single point of failure or misuse, they are less vulnerable.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Adhering to these principles will help improve data governance and contribute to trustworthy BDAS.

“Adhering to these principles will help improve data governance and contribute to trustworthy BDAS.”

Probably not a substantial claim: may not pass S2 (checkable).

As BDAS often depend on data sources that are external to the organization, there is a need to extend such data governance to cover multiple organizations.

“However, as BDAS often depend on data sources that are external to the organization, there is a need to extend such data governance to cover multiple organizations.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Critical aspects of data governance are identity and trusted data sharing, which are not part of the standard Internet infrastructure.

“Critical aspects of data governance are identity and trusted data sharing, which are not part of the standard Internet infrastructure.”

Probably not a substantial claim: may not pass S4 (based).

Apart from sociotechnical measures focused on controls and safeguards, an organizational culture that promotes awareness and ethical value of data and algorithms is part of data governance.

“Apart from sociotechnical measures focused on controls and safeguards, an organizational culture that promotes awareness and ethical value of data and algorithms is part of data governance.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

BDAS face the risks of violating the privacy, using data for undesired purposes, allowing bias or discrimination in data to inform algorithmic decisions, making wrong decisions, and so on.

“BDAS face the risks of violating the privacy, using data for undesired purposes, allowing bias or discrimination in data to inform algorithmic decisions, making wrong decisions, and so on.”

Classified as probably not a substantial claim. May combine two claims.

This raises security risks for identity information to be stolen, privacy risks with handling personal information, and data integrity risks as data might not be consistent with each other and the states of the real world.

“This raises security risks for identity information to be stolen, privacy risks with handling personal information, and data integrity risks as data might not be consistent with each other and the states of the real world.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

For a single organization, these types of decisions are made within the framework of data governance.

“For a single organization, these types of decisions are made within the framework of data governance.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Interoperability is created to be able to share data based on commonly agreed standards specified by the trusted framework.

“Interoperability is created to be able to share data based on commonly agreed standards specified by the trusted framework.”

Probably not a substantial claim: may not pass S4 (based).

BDAS often depend on data sources external to the organization, which can be easily manipulated or misused.

“BDAS often depend on data sources external to the organization, which can be easily manipulated or misused.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

When the data is hacked or shared for illegitimate purposes, this is immediately revealed.

“Then, when the data is hacked or shared for illegitimate purposes, this is immediately revelated.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

This also results in non-repudiation, that the data origin cannot be disputed.

“This also results in non-repudiation, that the data origin cannot be disputed.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

SSI provides control over and ownership of data to citizens, who can give consent to share such data with others.

“SSI provides control over and ownership of data to citizens, who can give consent to share such data with others.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The first situation entails organizations defining base registries using data stewardship.

“The first situation entails organizations defining base registries using data stewardship.”

Classified as probably not a substantial claim.

This results in fragmented data landscape where such information is stored in many places.

“This results in fragmented data landscape where such information is stored in many places.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Although these principles might at first appear simple, they are challenging to realize.

“Although these principles might at first appear simple, they are challenging to realize.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Although greater transparency into the inner workings of algorithms is necessary, it is insufficient for effective oversight.

“Although greater transparency into the inner workings of algorithms is necessary, it is insufficient for effective oversight.”