iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Universities can reduce AI surveillance without accepting more cheating by matching each educational purpose to the least intrusive method that can achieve it. Use human assessment or human proctoring for high-stakes decisions, redesign exams so they do not require room scans or continuous recording, limit learning analytics to the minimum data needed, add privacy-enhancing computation where aggregate insight is sufficient, and make every high-risk EdTech purchase conditional on transparency, human review, deletion and appeal rights.
What counts as AI surveillance in a university?
AI surveillance is any automated or semi-automated collection, inference or decision process that observes students or predicts their behaviour. Common examples include remote-proctoring software that records a room, face or keystrokes; biometric identification; gaze, emotion or “suspicion” inference; browser and device monitoring; and individualized learning-analytics scores that label a student as likely to fail, cheat or disengage.
The concern is not limited to whether a system uses a camera. A prediction built from log-ins, location, writing patterns or interaction histories can affect a student just as a recording can. The relevant questions are what is collected, what is inferred, who can see it, how long it remains available, and whether a person can challenge the result.
Recommended Free Tools
A privacy-preserving program therefore starts with the educational purpose and its threat model. A university should state the problem in one sentence—for example, “verify authorship of a final assessment” or “identify courses with unusually low participation”—then ask what the least data-intensive method would be. Protecting against organized contract cheating is a different problem from helping a student who has stopped attending, and neither automatically justifies continuous individual monitoring.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
How the alternatives compare
No single replacement fits every course. The choice should be documented against privacy intrusion, educational validity, equity and accessibility, contestability, implementation effort and the quality of evidence supporting the tool.
| Approach | Privacy exposure | Educational validity | Equity and accessibility | Contestability | Implementation considerations |
|---|---|---|---|---|---|
| In-person or human-proctored assessment | Usually lower than continuous automated surveillance; records can be limited to attendance and submitted work. | Strong when the format matches the learning objective; oral, practical and portfolio formats can test understanding directly. | Requires accessible rooms, scheduling and accommodations; remote human options can help students who cannot attend campus. | High: an instructor or trained proctor can explain and reconsider a decision. | Needs trained staff, scheduling and a documented procedure; comparative cheating or cost rates are not established by the cited guidance. |
| Assessment redesign | Low when it avoids room scans, biometrics and continuous recording. | Open-book, project, portfolio and process-evidence assessments can assess application, reasoning and development over time. | Can support varied modes and assistive technology, but deadlines and group work need accommodations. | High if rubrics, drafts and feedback are retained and students can respond to concerns. | Requires course-design time and clear rubrics; there is no universal effect size for cheating reduction. |
| Purpose-limited learning analytics | Moderate to low when data are minimized, aggregated and access-controlled; higher when individual profiles drive action. | Useful for operational questions such as course-level participation or resource demand, not a definitive measure of learning or intent. | Models can reproduce unequal access or disability-related patterns unless tested by subgroup and interpreted cautiously. | Must include visibility of labels, an explanation, human review and a route to correct errors. | Define purpose, data fields, retention and access before collection; validate data and models. |
| Privacy-enhancing analytics | Can reduce disclosure through aggregation, differential privacy, local processing or distributed analysis, but does not remove all risk. | Preserves useful statistical questions when individual identification is unnecessary. | Requires checking that noise or aggregation does not hide support needs for small groups. | Depends on clear documentation of guarantees, limits and the person responsible for decisions. | Needs a threat model, technical expertise, privacy-budget or access decisions and ongoing testing. |
| Human-centred governance controls | Limits collection and secondary use through policy, contracts and review. | Improves reliability by keeping people accountable for consequential judgments. | Requires impact assessment and consultation with affected students and staff. | Provides appeals, explanations, correction and sunset decisions. | Must be built into procurement and reviewed whenever a vendor changes a model or processing arrangement. |
1. Replace automated proctoring with human assessment routes
For a high-stakes or remote examination, offer a human route rather than making an automated surveillance system the only way to participate. OECD guidance says that continued use of AI-enabled remote proctoring should include an alternative human proctoring option because students have different connectivity, living space and examination conditions at home. That principle applies even when a university believes the software is accurate: a student should not have to accept biometric or room monitoring to obtain a fair assessment.
In-person examinations
A supervised campus exam can verify identity and conditions without recording a student’s home. The university must provide accessible seating, extra time, assistive technology and alternative rooms where required. Keep attendance and incident records separate from unnecessary video, and publish who can access any retained material and when it will be deleted.
Scheduled oral examinations
An oral examination, conducted in person or by a live human examiner online, can test whether a student understands submitted work. Use a shared rubric, record only when justified and agreed under applicable policy, and allow an accommodation when speech, hearing or connectivity creates a barrier. Oral assessment is not a universal substitute: it requires examiner time and should be aligned with the learning outcomes.
Projects, portfolios and process evidence
Require drafts, version histories, design notes, citations, demonstrations or short reflective explanations when the objective is authorship or applied competence. These artifacts let an instructor examine development rather than infer misconduct from a facial expression, typing pattern or room scan. State in advance which process evidence is required and how it will be graded.
Open-book and take-home work
Questions that require analysis of a new case, comparison of sources or a defended design are harder to answer by copying a stock response. Use staged deadlines, individualized data or oral follow-ups where appropriate, while avoiding a requirement to install monitoring software. The sources cited here support human alternatives and oversight but do not establish a particular cheating rate or cost advantage for any format.
2. Redesign assessments so surveillance is unnecessary
Start with the learning objective, then choose the evidence that demonstrates it. If the objective is calculation, provide novel data and require workings. If it is clinical judgment, use a case discussion or practical demonstration. If it is research communication, grade a proposal, source log, draft and revision. A camera pointed at a student’s room is not evidence of any of these outcomes.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
- Write the outcome. Specify the knowledge or capability being assessed, not the behaviour you hope to observe.
- Choose observable evidence. Select a product, performance, explanation or sequence of drafts that directly demonstrates the outcome.
- Set a transparent rubric. Describe quality levels, permitted resources, collaboration rules and how concerns will be investigated.
- Build in authorship evidence. Use checkpoints, annotations, demonstrations or a brief conversation rather than continuous recording.
- Test accessibility. Check keyboard access, screen-reader compatibility, time allowances, language needs and alternatives for students who cannot use a required device or location.
- Publish the review path. Tell students who investigates suspected misconduct, what evidence is considered and how to submit a response or appeal.
This design can reduce incentives for contract cheating while avoiding the false certainty of an automated “suspicion” score. It also makes an instructor’s judgment visible and contestable.
3. Use learning analytics with a narrow, stated purpose
Learning analytics can support students without tracking every action. Jisc’s code calls on institutions to define objectives, identify necessary data, set access boundaries and interpretation rules, explain whether AI is used, validate data and models, avoid identification or re-identification where it is unnecessary, and retain qualified human judgment. Students should be able to understand, question and correct an AI-generated prediction.
Define the question before collecting data
“Which modules have a sudden drop in participation this month?” can often be answered with aggregate counts. “Which named student is likely to fail?” creates a much higher-risk profile and may not be valid without context. Record the purpose, decision owner, data fields, permitted uses and prohibited uses in a public register.
Collect the minimum necessary fields
Do not ingest camera feeds, precise location, private messages or unrelated browsing history merely because a vendor can accept them. For a course-level participation report, weekly counts may be enough; individual timestamps and device fingerprints are not automatically necessary. Remove direct identifiers or aggregate results when an individual does not need to be known.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Separate support from punishment
A prediction can trigger an invitation to speak with an adviser, but it should not silently impose a penalty, lower a grade, block registration or rank students. A person must check the context: employment, disability, caring responsibilities, unreliable internet and many other factors can produce the same data pattern. Document when a human reviewed the output and what evidence they considered.
Give students visibility and correction
Explain what data and labels are held about a student, the purpose of the model, its known limitations and the consequences—if any—of an alert. Provide a simple correction channel and a response deadline. Corrections should flow back to the source system and any downstream decision record.
Apply a retention schedule
Set deletion dates for raw events, derived features, alerts and audit logs. Retaining an old risk label indefinitely can affect a student after the original context has disappeared. Access should be limited to staff with a legitimate need, with logs reviewed for inappropriate use.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
4. Add privacy-enhancing computation carefully
When a university needs statistics rather than a person’s identity, technical controls can reduce exposure. NIST defines differential privacy as a mathematical framework that quantifies privacy risk to individuals from data collection and subsequent data release. A system may add calibrated noise or otherwise limit the contribution of any one person before publishing an aggregate.
Differential privacy
Differential privacy is useful for questions such as enrollment trends or resource demand where exact individual values are not required. The university must document the privacy guarantee, parameters, composition across repeated releases, small-group suppression rules and the utility loss caused by noise. NIST stresses evaluating guarantees and implementation hazards; differential privacy is one control, not permission to collect unlimited data or a substitute for governance.
Local and on-device processing
Where feasible, process an event on a student’s device and transmit only a result or aggregate. This can reduce central storage, but it does not automatically solve consent, security or model-bias problems. Verify what leaves the device, how updates are delivered and whether a vendor can access diagnostic data.
De-identification and re-identification testing
Removing names is not enough when combinations such as course, timestamp and device can identify a person. Test whether a reasonably capable party could reconnect records, restrict auxiliary data and treat the result as lower risk only when the test supports that conclusion.
Federated or distributed analysis
In a federated design, participating systems keep data locally and share model updates or statistics. This can reduce transfers, but updates may still leak information and governance becomes more complex. Choose it only after documenting the threat model, secure aggregation, update auditing and responsibility for incidents.
Role-based access and short retention
Technical architecture should enforce the policy: separate identifiers from analytics, grant least-privilege access, encrypt transfers and storage, alert on bulk exports, and delete temporary data on schedule. A privacy-enhancing design that no one monitors can still fail in practice.
5. Govern AI as a human-centred educational system
UNESCO’s guidance of 7 September 2023 calls for a human-centred approach, data-privacy protection, ethical validation and coherent policy frameworks for AI in education and research. OECD guidance likewise emphasizes transparency about objectives, functioning and limitations, timely human support and a human alternative where appropriate.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
A university-wide policy should require the following before a high-impact system is used:
- Purpose and necessity statement: the educational problem, expected benefit and reasons less intrusive methods are inadequate.
- Data inventory: every input, inferred label, recipient, storage location, retention period and secondary use.
- Legal and ethical basis: the applicable authority for processing, student notice, consent where relevant and an assessment of power imbalance.
- Risk and impact assessment: foreseeable privacy, security, discrimination, accessibility and academic-freedom harms.
- Validation evidence: performance, error types and subgroup testing in the university’s population and operating conditions.
- Human accountability: named decision owners, training, escalation rules and a ban on rubber-stamping automated outputs.
- Notice and appeal: plain-language explanations, access to records, correction, reconsideration and an independent appeal route.
- Sunset and renewal: a fixed review date and a decision to stop, modify or renew based on measured benefit and residual risk.
Document consultation with students, disability services, instructors, librarians, security staff, data-protection officers and unions where relevant. Consultation is not a replacement for a risk assessment; it reveals harms and practical constraints that a technical review can miss.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitches6. Procurement rules for EdTech vendors
UK Department for Education guidance offers a practical pattern for high-risk educational technology: use data protection by design and default, complete a data-protection impact assessment (DPIA), involve the data-protection officer (DPO) early, and require suppliers to explain model operation, training use, controls, inaccuracies and bias. Contracts should address subprocessors, security, deletion, return, breach notification, staff access and review when the tool changes.
Pre-contract checklist
- Can the university disable model-development or vendor-training use of its data?
- Is each data field and inferred attribute listed, with a documented purpose?
- Does the supplier state where data and backups are processed, including international transfers?
- Can the institution set retention periods and obtain verifiable deletion and return at termination?
- Are subprocessors disclosed in advance, with notice and objection rights for changes?
- What encryption, access logging, incident response and independent assurance are provided?
- What are the model’s known error rates, limitations, accessibility issues and subgroup results?
- Can staff export an audit trail showing the input, output, human action and appeal outcome?
- Can a student obtain an understandable explanation and correct inaccurate data?
- Does the contract prohibit using an automated output as the sole basis for a high-impact decision?
- What happens when the model, hosting location, subprocessor or terms change?
The DfE guidance records a case in which a school rejected an AI writing tool because it could not switch off model-development use and stored data outside the UK. For a university, inability to disable secondary use, explain processing, meet transfer requirements or support deletion and oversight is a procurement stop condition, not a minor contract issue.
Contract clauses to make explicit
Specify the university’s role and the supplier’s role; permitted processing instructions; confidentiality and staff-access limits; security controls; breach-notification deadlines; subprocessor approval; international-transfer safeguards; assistance with rights requests and DPIAs; deletion and return formats; audit rights; insurance and liability; and termination if the supplier cannot meet the requirements. Require advance notice and a fresh review for material model or data-flow changes.
7. A staged implementation plan
Stage 1: Pause unjustified surveillance
Freeze new deployments and any use that lacks a documented purpose, impact assessment, human alternative or retention rule. Preserve only records needed for current academic-integrity cases, and stop secondary uses that are not necessary for the stated purpose.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Stage 2: Map systems and consult affected communities
Create an inventory of cameras, microphones, biometric features, browser telemetry, learning-management events, predictive labels, vendors and subprocessors. Ask students and staff which effects are experienced, including accessibility and housing constraints. Publish a plain-language summary.
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
Stage 3: Pilot the least intrusive method
For each use case, pilot human assessment, redesigned assessment or aggregate analytics before individual monitoring. Define success measures in advance: learning-objective validity, completion and support outcomes, accessibility incidents, false alerts, appeal outcomes, staff workload and privacy events.
Stage 4: Measure disparate impacts and unintended effects
Compare error and burden patterns across relevant groups where lawful and statistically responsible. Investigate whether a tool disadvantages students with disabilities, limited connectivity, shared housing, caring duties or different language backgrounds. Do not treat a lack of complaints as proof of fairness.
Stage 5: Publish results and decide
Report the purpose, data used, safeguards, observed benefits, limitations, incidents and student feedback. Renew only when the measured educational benefit justifies the residual risk and the human alternative remains available. Otherwise retire the system and delete data according to the schedule.
Common failure modes and fixes
- “Human in the loop” means a person clicks approve. Give the reviewer time, evidence, training and authority to disagree; record the reasoning.
- Anonymization is assumed to be permanent. Test re-identification and reduce granularity, access and retention.
- Consent is treated as a complete solution. Students may have no realistic alternative in a compulsory course; provide a comparable human route and assess necessity.
- Analytics alerts become discipline scores. Restrict use to the stated support purpose and prohibit penalties or rankings unless a new, justified governance decision permits them.
- Vendor updates bypass review. Contractually require notice and reassess when models, subprocessors, hosting or training uses change.
- Privacy is measured only by storage volume. Include inference, access, transfer, retention, re-identification and the consequences of an error.
- Small groups disappear in aggregates. Apply suppression or carefully evaluated differential privacy, while preserving a safe human route for support.
What universities should require before deployment
- A specific educational purpose and a documented least-intrusive alternative.
- A DPIA and DPO review for high-risk processing, completed before procurement or pilot use.
- A data map covering inputs, inferences, recipients, locations, subprocessors and retention.
- Independent validation, subgroup and accessibility testing, and a description of known limitations.
- A human decision-maker with authority to override the system and a timely appeal process.
- Student notice that explains data, model use, consequences and correction rights.
- Contractual controls for training use, transfers, security, deletion, breach response, audit and updates.
- Success and harm metrics, public reporting and a scheduled sunset or renewal decision.
Frequently asked questions
Can a university stop AI proctoring without increasing cheating?
It can stop making automated proctoring the default and use human proctoring, in-person exams, oral checks, projects, portfolios and process evidence matched to the learning objective. The cited guidance supports these alternatives and human oversight, but it does not provide a universal cheating-rate comparison; each institution should measure its own outcomes.
Is differential privacy enough to make student analytics safe?
No. It quantifies and limits certain disclosure risks in data releases, but privacy also depends on purpose, access, retention, security, inference, contracts and human decisions. NIST recommends evaluating guarantees and implementation hazards as part of a broader program.
Should individual risk predictions be banned?
They should not be used casually or as automatic grounds for penalties, ranking or restrictions. If a university considers an individual alert for support, it needs a defined purpose, minimum data, validation, subgroup testing, human review, student explanation and correction route. A course-level aggregate may answer the same question with less risk.
When is a DPIA required?
Use a DPIA before high-risk processing, especially biometric or intensive profiling, and involve the DPO early. The exact legal trigger depends on the university’s jurisdiction, but waiting until after deployment defeats the assessment’s purpose.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWhat if a vendor refuses to disclose its model?
Do not treat secrecy as a reason to weaken oversight. Require enough information to assess purpose, inputs, training use, accuracy, bias, security, transfers, retention and appeal. If the supplier cannot support explanation, deletion or independent review, choose a less intrusive method or do not procure it.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
