Complete underlying reasoning (54 claims and questions)
The main steps above are a reading path. These are all assessed contributions to this argument, including challenges and questions.
Answered · How does the article organize its argument and conclusion?
The roadmap says resistance is morally undesirable and satisfactory accommodation is difficult; the conclusion says governance may retain instrumental and procedural virtues while sacrificing human control and understanding. This materially connects to the narrative’s trade-offs, while remaining a summary of the argument rather than its full development.
How does the article organize its argument and conclusion?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesPartly answered · Why is it hard to keep people involved in decisions made by algorithms?
The question concerns accommodation’s difficulty and is relevant to the narrative’s resistance/accommodation trade-offs. The supplied answer only distinguishes kinds of responses; it leaves the reasons accommodation is difficult unresolved.
Why is it hard to keep people involved in decisions made by algorithms?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Is resisting systems that use algorithms always the morally better choice?
The narrative records costs and benefits of resistance and accommodation without treating either as a demonstrated solution.
Is resisting systems that use algorithms always the morally better choice?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · What practical value might systems that use algorithms bring to large collections of data?
This possible benefit is relevant to the narrative’s balancing of practical gains against participation and understanding risks.
What practical value might systems that use algorithms bring to large collections of data?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Could people resist systems that use algorithms in all areas of public decision-making?
The narrative includes resistance’s practical difficulties and trade-offs.
Could people resist systems that use algorithms in all areas of public decision-making?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered under conditions · When might the author consider it reasonable to rely on an algorithm?
Raises a conditional benefit relevant to weighing resistance against accommodation; it does not establish a general solution.
When might the author consider it reasonable to rely on an algorithm?
Artificial intelligence browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Can an algorithmic system bring practical gains in how work gets done, even when people disagree about the wider workplace practices around it?
Raises a possible operational benefit relevant to weighing resistance, without resolving wider concerns or establishing that the benefit settles them.
Can an algorithmic system bring practical gains in how work gets done, even when people disagree about the wider workplace practices around it?
Artificial intelligence browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · How might people accept control by algorithms, or rules a computer follows, without understanding their reasons?
Raises how promised benefits may encourage accommodation despite reduced clarity, a trade-off rather than a demonstrated outcome.
How might people accept control by computer systems without understanding their reasons?
Question created with artificial intelligence browsing, based on this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · How could a system that uses computer rules and first seems useful become a threat?
Raises how perceived benefits may encourage reliance while understanding declines, relevant to the narrative’s trade-off.
How could a system that uses computer rules and first seems useful become a threat?
Question created with artificial intelligence browsing, based on this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Could it become harder over time for people to notice when systems collect data about them?
Visibility and ubiquity bear on the narrative’s account of resistance difficulties and trade-offs; this is a forecast, not a demonstrated trend.
Could it become harder over time for people to notice when systems collect data about them?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesPartly answered · How should concerns about systems that use algorithms be weighed against their benefits?
Opens the issue of benefits versus risks to understanding and participation; the answer identifies the need to weigh them but does not settle how to do so.
How should concerns about systems that use algorithms be weighed against their benefits?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Why might resisting algorithmic decision systems be difficult?
Asks about the resistance side of the narrative’s trade-off; the answer gives the author’s reasons, not a demonstrated general outcome.
Why might resisting algorithmic decision systems be difficult?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · What possible gains and losses from automated decision systems should be compared?
This is the benefits-versus-costs frame in the narrative: possible system benefits alongside risks to participation and understanding.
What possible gains and losses from automated decision systems should be compared?
Artificial intelligence browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered under conditions · Could algorithmic systems improve how decisions are made, not just their results?
The narrative includes possible procedural benefits alongside risks to participation and understanding; the answer’s support remains limited because the cited argument is not explained in the supplied passage.
Could algorithmic systems improve how decisions are made, not just their results?
Artificial intelligence browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Does the author favor rejecting automated decision systems or finding ways to keep people involved?
The question concerns the resistance-versus-accommodation trade-off described in the narrative.
Does the author favor rejecting automated decision systems or finding ways to keep people involved?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Could giving people less room to use their own judgment make decisions fairer to groups with less power?
The question raises a possible fairness benefit of algorithmic systems that is relevant to the narrative’s weighing of procedural benefits against risks to participation and understanding. It remains Zarsky’s argument, not Danaher’s endorsed conclusion or an established result.
Could giving people less room to use their own judgment make decisions fairer to groups with less power?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · What could happen if no remedy is available?
The answer describes a possible participation cost and trade-off if no effective response is available.
What could happen if no remedy is available?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Why might some people prefer people to decide whom to profile, rather than use automated profiling?
The attributed explanation concerns resistance to automated profiling, but is not Danaher’s endorsed account.
Why might some people prefer people to decide whom to profile, rather than use automated profiling?
Artificial intelligence browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Could AI partnerships turn people from decision-makers into recipients of AI guidance?
Possible loss of agency is a cost relevant to accommodating algorithmic systems.
Could AI partnerships turn people from decision-makers into recipients of AI guidance?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Can systems that use algorithms support the use of renewable energy?
The example concerns a possible practical benefit relevant to weighing benefits against concerns; it does not resolve that broader trade-off.
Can systems that use algorithms support the use of renewable energy?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Could working with AI help people remain employable as AI advances?
The reported employability claim is relevant context about a possible practical benefit of accommodation; the chess analogy is explicitly qualified and does not establish that benefit for work.
Could working with AI help people remain employable as AI advances?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · What does the author expect from efforts to respond to the threat?
This directly concerns the trade-offs of resistance and accommodation; the answer does not establish a solution.
What does the author expect from efforts to respond to the threat?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesAnswered · Why might people find it difficult to resist systems that collect data about them?
This directly concerns practical obstacles to resisting algorithmic systems; it reports the author’s argument, not a tested intervention.
Why might people find it difficult to resist systems that collect data about them?
AI browsing question drawn from this source idea; not a quotation or a claim that the authors asked this question.
Inspect the contribution and sourcesImportant limit · Weighing possible gains in results and fairness against losses in understanding and participation
This specifies dimensions relevant to the costs and benefits of resistance and accommodation.
The author says automated decision systems might produce better results and make the decision process fairer. These possible gains must be weighed against the possibility that people may understand less or have less say in decisions.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. The author frames the choice as a balance among possible results, procedural fairness, understanding, and participation.
Inspect the contribution and sourcesImportant limit · The intended benefits of partnering with algorithms
These intended benefits are relevant to accommodation’s possible attractions but remain proposals subject to the author’s doubts.
The intended case for partnering with algorithms is that people could keep their independence, strengthen their thinking and understanding, and remain able to take part meaningfully in decisions shaped by algorithms.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. These are the proposal’s intended benefits; the author then questions whether the proposal can deliver them. Addresses: The proposed purpose of partnering with algorithms.
Inspect the contribution and sourcesImportant limit · Convenience may hide how a system controls people
It describes a possible exchange of control and clarity for promised benefits, relevant to accommodation’s trade-offs, not a demonstrated general outcome.
The paper uses Morozov’s warning to describe systems that can steer people toward or away from actions without explaining why. People may accept this control in return for promised personal or social benefits, while supplying the system with their data and receiving little clarity about how it produces those benefits.
Source coverage: Full readable text of the author’s manuscript from an accessible mirror; this is not the final version formatted by the publisher. This is a view the author quotes and discusses, not a finding from a study of users in these passages. Addresses: Morozov’s argument as presented in the paper
Inspect the contribution and sourcesImportant limit · People can build bias into a system as they design it
It limits the possible fairness benefit of automation: careful attention may reduce design-stage bias, but the supplied claim does not say it eliminates bias.
Automated systems can repeat bias because people decide how to turn a task into instructions and code, and bias can enter at either step. Paying close attention to this risk may help reduce bias, but the passage does not say it can be eliminated.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. The passage explains both a route by which bias can enter and a possible way to limit it.
Inspect the contribution and sourcesImportant limit · Algorithms may help people use large collections of data
Potential benefits of algorithmic use are relevant to the narrative’s claim that resistance and accommodation involve trade-offs, not a demonstrated solution.
The author argues that algorithmic systems may help public bodies and the public use growing collections of data.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage says that increasingly large datasets may be put to good use through algorithmic technologies, for both social authorities and the public. Addresses: The possible benefits of using large datasets
Inspect the contribution and sourcesImportant limit · Some algorithmic systems predict better in some fields
Possible performance benefits bear on trade-offs about resistance, but the claim is limited and does not resolve the trade-off.
The author reports that earlier research found algorithmic systems sometimes made better predictions than human experts. The author does not claim this was true in every field or show that tax enforcement was one of those fields.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. The passage cites earlier work for the claim that algorithms outperform human experts on predictions in certain fields. It uses this as support for possible deference in some cases. Addresses: The scope of reported prediction findings
Inspect the contribution and sourcesImportant limit · Fairness and relevant reasons for profiling
It supplies a possible fairness benefit that may be weighed against risks, while warning that reduced bias is not guaranteed.
The author describes a possible fairness benefit: automated profiling might avoid some bias that people may not notice in their own decisions. But the concern is not profiling itself. It is targeting people arbitrarily rather than on legitimate grounds, meaning relevant reasons for targeting them, such as being genuinely more likely than others to commit the crime in question. The author presents reduced bias as a possibility, not a guarantee.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. Passage 25 distinguishes arbitrary discriminatory targeting from targeting on legitimate grounds, meaning relevant reasons for targeting, then describes a possible advantage of automated systems. Passage 26 adds that automated systems can also reproduce bias.
Inspect the contribution and sourcesImportant limit · The author weighs reasons to resist against reasons to use these systems
Qualifies opposition to algorithmic systems by noting possible benefits that must be weighed against participation and understanding risks.
The author argues that resistance is not automatically the morally better choice, because algorithmic systems may have practical benefits and the way they make decisions may itself provide a reason to use them.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage says there are moral arguments for using such systems based on their outcomes and on the way decisions are made. It therefore frames resistance as involving competing values. Addresses: The moral case for and against resistance
Inspect the contribution and sourcesSupporting point · Resistance has costs
Supports the narrative’s claim that resistance and accommodation both carry risks and costs, without establishing a solution.
The author argues that simply stopping the use of algorithmic decision systems may be difficult or undesirable because these systems are becoming more common and less noticeable, while also offering practical benefits and benefits in how decisions are made.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The author says the costs of opacity must be weighed against the systems’ other benefits.
Inspect the contribution and sourcesImportant limit · Human participation may shrink
Highlights the stakes of unresolved resistance and accommodation: participation may shrink even where other benefits are pursued.
The author warns that society may create governance systems that sharply restrict people’s chances to take part, with no readily available remedy; such limits might bring other benefits, but people need to decide whether that trade-off is acceptable.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The conclusion presents this as a serious possible cost of pursuing other practical or procedural gains.
Inspect the contribution and sourcesImportant limit · Look for ways to preserve participation
This favors considering accommodation rather than simple rejection, without claiming that preserving participation alongside use solves the trade-off.
Because the trade-off is complex, the author questions whether rejecting automated decision systems is the right response. He suggests keeping possible benefits while finding other ways for people to participate.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The author presents preserving participation alongside continued use as a possibility, rather than a settled solution.
Inspect the contribution and sourcesSupporting point · Smart electricity grids may help renewable energy
The example supplies a possible practical benefit relevant to weighing resistance against accommodation; it does not establish the overall balance of benefits and risks.
The author gives smart electricity grids as an example of data-monitoring systems that can help renewable energy work more effectively and efficiently.
Source coverage: Full readable text of the author’s manuscript (the version written by the author), found on an accessible mirror; not the final version prepared by the publisher. The passage says smart grids rely heavily on monitoring and data mining (finding patterns in data), and presents them as a way to improve both the effectiveness and the efficiency of renewable energy sources. Addresses: The smart-grid example
Inspect the contribution and sourcesImportant limit · Political action and deliberate damage or disruption may be hard to carry out
Identifies a practical limit on resistance, without concluding that resistance is impossible or that accommodation is preferable in every case.
Calls to resist systems that use algorithms through political action or sabotage—deliberately damaging or disrupting them—face a practical challenge: the author doubts such resistance could work across all areas where public decisions are made.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage reports Morozov’s call to make the issue political and to sabotage systems by deliberately damaging or disrupting them, then questions whether this kind of resistance can work across all areas where public decisions are made. Addresses: The practical reach of resistance
Inspect the contribution and sourcesSupporting point · Rules may face resistance
Shows a potential cost and source of opposition relevant to balancing accommodation against other responses.
Requiring artificial intelligence (AI) decision systems to be understandable may reduce some of their practical benefits and face resistance from businesses or governments that benefit from them.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The author says interpretability requirements could reduce the benefits sought from these systems and meet resistance from interested organizations. Addresses: Whether human review can be required in practice
Inspect the contribution and sourcesImportant limit · Partnership remains possible
Shows that accommodation is not categorically rejected, while making clear that partnership faces unresolved difficulties.
The author does not rule out people working with technology as part of a response, but says this approach faces both technical and philosophical difficulties.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. This qualifies the earlier criticism: the author leaves room for partnerships while emphasizing their difficulties. Addresses: The proposals for closer human–technology partnership.
Inspect the contribution and sourcesImportant limit · Team success does not ensure human understanding
Adds a risk of diminished human understanding to proposed accommodation through partnership; it does not show that all partnerships have this result.
The author says it is unclear whether personal AI partnerships would let people take part in decisions in a meaningful way. The chess example raises a concern: the strongest teams were not those whose human members understood the game best, and top players could be hindered by second-guessing the computer.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The chess example is used to question whether better joint performance means the human partner understands or guides the process. Addresses: The concern that AI partnerships may not support meaningful human participation.
Inspect the contribution and sourcesImportant limit · Deep integration may not be possible now
Limits the viability of one accommodation proposal while leaving other possible responses unresolved.
The author objects that deep human–technology integration is highly speculative, may not be conceptually possible, and is far from practical realization.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. These objections directly challenge the feasibility of the proposed response. Addresses: The proposal to integrate people and technology at the level of thought.
Inspect the contribution and sourcesImportant limit · The author favors overcoming expert reliance
Recognizes existing expert reliance as a compromise, but does not treat its existence as proof that it should be preserved.
The author treats reliance on experts in public decisions as a compromise between useful results and public participation. He says that if this compromise could be overcome, it probably should be.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage acknowledges that people already depend on experts, but says this practice is not necessarily worth preserving and that overcoming it would probably be preferable if possible. Addresses: The argument that expert dependence already exists
Inspect the contribution and sourcesWhat this suggests · Both responses have limits
States the narrative’s central trade-off while preserving that these are the author’s assessments, not tested intervention findings.
Danaher says neither resisting the threat nor finding a satisfactory way to live with or manage it is likely to succeed without risking other things people value about social decision-making. He separately describes resistance as morally undesirable and says that finding a satisfactory way to live with or manage the threat would be difficult.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the version formatted by the publisher. These are the author’s assessments of proposed responses, not findings from approaches that have been tested; the supplied passages do not specify the risks or give the full reasoning. Addresses: The author’s assessment of resistance and accommodation (finding a satisfactory way to live with or manage the threat)
Inspect the contribution and sourcesImportant limit · A reported reason to prefer human profiling
Adds an attributed, non-endorsed account of a possible motivation for resistance; it does not establish that motivation as true.
The author reports Zarsky’s argument that people in a privileged majority may prefer human-run profiling because human bias could favor them over minorities. Danaher says he does not endorse Zarsky’s argument.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. This is a position attributed to Zarsky, and the author explicitly distances himself from endorsing it.
Inspect the contribution and sourcesImportant limit · Relying on an algorithm depends on the outcome
Supplies a conditional instrumental reason to use algorithmic systems without resolving their participation costs.
The author’s case for relying on an algorithm is conditional: it may make sense in some cases if doing so leads to better outcomes.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. The passage says that in many cases people may be well advised to follow an algorithm’s recommendation if they want better outcomes. It does not argue for rule by AI or for always following algorithms. Addresses: Conditional support for algorithmic decisions
Inspect the contribution and sourcesWhat this suggests · The difficult choice between responses
The stated conclusion juxtaposes instrumental and procedural virtues with possible losses of human control and understanding; this is a roadmap-level summary, not a tested intervention.
The article lays out a three-part argument: it makes the case for the threat, argues that resisting it is morally undesirable, and argues that satisfactory accommodation is difficult. Its conclusion says governance may retain useful features and good ways of making decisions while sacrificing human control and understanding.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the version formatted by the publisher. The supplied passages give this roadmap and conclusion, but not the full arguments developed in the later sections. Addresses: The article’s roadmap and stated trade-off
Inspect the contribution and sourcesImportant limit · Algorithms may also improve how decisions are made
Adds a possible countervailing benefit in decision procedures while preserving the source’s limited detail and uncertainty.
The author notes that algorithmic systems may offer benefits in how decisions are made, as well as in their results. The author points to Zarsky’s argument, but the supplied passage does not explain these benefits in detail.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. The passage distinguishes benefits in outcomes from benefits in decision procedures and refers to Zarsky’s case for the latter. The text provided ends before explaining that case. Addresses: Possible procedural benefits
Inspect the contribution and sourcesImportant limit · Monitoring may become more widespread
Adds a possible reason resistance may become harder as monitoring grows less visible, preserving the source’s predictive uncertainty.
The author expects monitoring technologies to become smaller, more efficient, and more widespread, which could make them harder for people to notice.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage says monitoring is becoming more common and harder to notice, and suggests these trends are likely to continue as monitoring technologies change. Addresses: The possible future growth of monitoring technologies
Inspect the contribution and sourcesSupporting point · How the system stores goods
The reported reduction in waste and faster stock turnover illustrate practical benefits relevant to weighing resistance against accommodation, without establishing that benefits outweigh risks.
The warehouse system uses available shelf space to assign goods to locations, rather than grouping similar goods together. The author says this can mean less wasted product and faster stock turnover, while placing unlike products beside one another.
Source coverage: Full readable text of the author’s manuscript from an accessible mirror; this is not the final version formatted by the publisher. This is an illustration of a reported system, not a result from a study conducted by the author.
Inspect the contribution and sourcesImportant limit · Weigh risks against benefits
The risk-benefit framing reinforces the narrative’s account of trade-offs without resolving which response is preferable.
The author says the risks of systems that use algorithms, including difficulty seeing how they work and possible effects on privacy, should be weighed against their benefits rather than considered alone.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage argues that difficulty seeing or understanding how a system works is one concern among several and says it should be considered alongside privacy effects and possible benefits. Addresses: Balancing risks and benefits
Inspect the contribution and sourcesSupporting point · Warehouse algorithms may reduce waste
It gives a bounded example of practical benefits relevant to weighing resistance against accommodation; it does not settle that balance.
The author says Amazon’s warehouse algorithms can reduce waste and inefficiency and increase profitability, while noting that this does not settle broader questions about the company’s work practices.
Source coverage: The full text of the research paper written by the author, read from a copy hosted on another site; not the final version formatted by the publisher. The passage credits the algorithms with reducing waste and inefficiency and increasing profit, while expressly separating that claim from judgments about the company and its broader work practices. Addresses: The claimed benefits of Amazon’s warehouse algorithms
Inspect the contribution and sourcesObjection · A possible protection from human bias
The attributed, tentative procedural-fairness argument is a counterconsideration in weighing potential benefits against participation and understanding risks; it does not establish that algorithmic decisions are fairer.
Zarsky argues that reducing the room people have to use their own judgment could protect minorities and other groups with less power from decisions shaped by human bias. Danaher presents this as a possible benefit of the decision-making process, not as an established result, and says he does not endorse Zarsky’s argument.
Source coverage: The full readable text of the author’s draft, from a website hosting a copy; not the final version formatted by the publisher. Zarsky argues that decisions led by computer systems can protect groups with less power by limiting the room people have to use their own judgment. Danaher does not endorse this argument.
Inspect the contribution and sourcesSupporting point · Good results may lead people to give a system’s judgment more weight
The appeal of apparent benefits helps explain why resistance and accommodation involve trade-offs.
The paper argues that decision systems based on algorithms—a set of rules a computer follows—may be especially likely to become hard to understand. People might first favour them for apparent gains in speed, accuracy, and ability to see patterns or draw conclusions, then come to rely on them without being able to understand how they work.
Source coverage: Full readable text of the author’s manuscript from an accessible mirror; this is not the final version formatted by the publisher. The author presents a possible gradual path from a system’s apparent usefulness to growing difficulty understanding how it works, not a finding that every such system follows it. Addresses: The paper’s account of the risk of rule through algorithms
Inspect the contribution and sourcesSupporting point · Monitoring can be hard to notice
Everyday reliance and low visibility are specific obstacles to resistance, not proof that resistance is impossible.
Widespread data monitoring can make resistance difficult. People may rely on these technologies in daily life and become less aware of how they are used and what effects they have.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The passage argues that monitoring technologies and technologies that analyze collected data are widespread and often hidden in plain sight. It says daily reliance can make people less aware of their possible role in systems that use algorithms to make or guide decisions. Addresses: The everyday presence and low visibility of monitoring
Inspect the contribution and sourcesImportant limit · Keeping people involved is difficult
It supports the account of accommodation as difficult, while distinguishing the specific problem of opaque algorithmic control from expert rule.
Keeping people meaningfully involved while retaining the benefits of algorithmic systems is difficult. Some proposed responses may be unrealistic. Others address the separate problem of leaving decisions to a small group of knowledgeable people, rather than the problem of decisions being controlled by algorithms that people cannot understand.
Source coverage: Full readable text of the author manuscript from an accessible mirror; not the publisher typeset version. The author distinguishes the threat of government decisions controlled by algorithms from the different risk of leaving decisions to a small group of knowledgeable people.
Inspect the contribution and sources