Singapore University of Technology and Design (SUTD) Information Systems Technology and Design (ISTD) latest news in November 2026 highlights an urgent issue for Singapore businesses: a Cyber Security Agency of Singapore survey found that 55% of 2,005 surveyed organisations had adopted AI, but those surveyed had implemented an average of only 39% of relevant basic AI-related cybersecurity measures. The School also reported research helping autonomous vehicles understand surroundings in rain or darkness, and a student-linked AI assistant called Sunday, which won a S$16,000 top prize for helping visually impaired users perform smartphone tasks.

For students preparing for SUTD 2027 admission, the most important distinction is that the ISTD pillar’s undergraduate degree is now called Bachelor of Engineering (Computer Science and Design, CSD). It was previously named Information Systems Technology and Design. The current degree spans eight academic terms and offers Computing-intensive pathways in Artificial Intelligence, Security, Software Engineering, Data Analytics, Fintech, IoT and Visual Computing. On 12 November 2026, SUTD’s official undergraduate admissions pages still showed 2026, not confirmed 2027, application dates. This independent eduKateSingapore guide therefore avoids inventing next year’s closing deadline.

Quick answers: SUTD ISTD in 2026 and 2027

• 10 November 2026: SUTD reported commentary by Associate Professor of Practice Victor Keong on businesses adopting AI faster than they implement adequate cyber safeguards.
• CSA Singapore Cybersecurity Health Report 2025: 55% of 2,005 organisations surveyed used AI, but had put in place 39% of applicable basic AI cybersecurity measures on average.
• 7 November: the School highlighted cyberattacks and data breaches affecting essential services and supply chains, warning about weak incident preparation.
• 4 November: Associate Professor Daisuke Mashima discussed risks to essential services and personal data in a public broadcast.
• 28 October: Sunday, a smartphone AI assistant intended to support people with visual impairments, won a S$16,000 top award at Huawei’s Tech4City competition.
• 6 October: Assistant Professor Zhao Na led research on balancing cameras and LiDAR so self-driving perception models work better in rain, darkness and unfamiliar cities.
• 24 September: the ISTD research team reported IterFlow, a lightweight framework helping 4D radar estimate three-dimensional motion in traffic.
• 21 September: a theory-guided AI network-access system was reported to halve congestion-clearing time under its research comparisons.
• 10 September: the School highlighted HieraScaffold, a framework for generating more realistic dynamic 4D LiDAR simulation scenes.
• Current undergraduate degree: BEng Computer Science and Design, not a newly admitted degree named BEng ISTD.
• Curriculum: the official ISTD page lists CSD core, technical electives, Humanities and Social Sciences and an interdisciplinary capstone, after Freshmore.
• AY2026 subsidised tuition: S$13,500 annually for qualifying Singapore citizens under SUTD’s cohort-based fee model, not a guaranteed 2027 price.
• Next action: verify qualification-specific SUTD 2027 dates when issued, then compare CSD with Design and AI, EPD and related Computing courses.

10 November: businesses adopt AI faster than their safeguards

The SUTD news article dated 10 November 2026 summarises reporting on the Cyber Security Agency of Singapore’s Singapore Cybersecurity Health Report 2025.

The survey included 2,005 organisations. SUTD’s summary states that 55% had adopted AI in business operations, while the surveyed organisations had implemented an average of 39% of basic AI-related cybersecurity measures.

The university also reported that nearly 90% of surveyed organisations experienced at least one cybersecurity incident during the previous year, but explicitly noted that not all those incidents were caused by or related to AI.

This is critical for truthful interpretation. It would be wrong to present the figures as proof that 90% of AI systems were hacked, or that each AI adopter had exactly 39% of its measures implemented.

What is Shadow AI?

Shadow AI describes employees using AI tools without appropriate organisational knowledge, approval or oversight.

An employee might paste confidential customer information into a public chatbot to produce personalised emails or upload internal code for debugging.

The employee may be seeking to complete work faster, but they can create privacy, contractual or intellectual-property risks if the service’s terms and organisational policies do not permit that data use.

Associate Professor of Practice Victor Keong emphasised that AI can accelerate work but can also increase the speed at which mistakes happen.

A sound company policy should help people understand which tools are approved, which data may be used and how results must be checked.

Why “we use AI” is not a security plan

A business might purchase an AI subscription and assume the provider is responsible for every aspect of data protection.

In practice, responsibility can be shared. The organisation controls who has access, what data employees submit and whether an AI agent receives permission to send messages or change records.

Basic measures such as multi-factor authentication, software updates, access controls, backup and activity logs can be valuable.

A high price does not automatically make a system secure. The relevant safeguards must match the actual risks.

A practical example: a customer-list upload

Imagine a staff member copies a customer spreadsheet into an external AI tool and asks it to write marketing messages.

The spreadsheet contains names, contact information and purchase histories. Even if the tool produces effective text, the upload may breach a policy or create privacy risk.

A better process would determine which information is genuinely needed, apply authorised data handling and use approved systems.

A Security student should understand these controls before recommending any AI workflow.

The point is not to ban all AI in business. It is to make technology use accountable and lawful.

What happens when an AI agent has excessive permissions?

An AI agent may carry out connected steps, such as reading messages, creating a report, sending emails or updating a database.

A mistaken instruction could then move beyond producing an incorrect sentence to performing an incorrect action.

For example, an agent with permission to delete records might misunderstand a housekeeping request and erase information that the organisation needed to retain.

The safest system defines clear permissions, checks before consequential actions, preserves logs and gives people meaningful oversight.

This is why secure computing and software design remain important even as AI models improve.

7 November: cybersecurity threats to supply chains and essential services

The ISTD news archive records a 7 November 2026 discussion about cyber incidents affecting European cold-chain logistics and local concerns about data breaches.

Associate Professor Keong warned that many small and medium-sized enterprises may have limited security resources and inadequate incident-response plans.

A business that provides software or services to a larger organisation can become a point of vulnerability in the supply chain.

The problem is not only financial theft. Disruptions affecting logistics, healthcare or other essential functions can have wider consequences.

Why cybersecurity is a systems problem

An organisation may protect its central servers but overlook a supplier’s account, a connected device or an employee’s outdated application.

A weakness in one part of the system may affect other services that depend on it.

Security therefore involves identity, software patching, networks, backups, employee training and contractual responsibility.

A cyber incident may require cooperation among technical teams, legal advisers, management and outside authorities.

No single product prevents every possible incident.

A student exercise: building a basic security model

Suppose a school event uses a small website for participant registration.

Students can draw a diagram showing where form data goes, who can view it, how long it is needed and what happens if the server becomes unavailable.

They can identify risks without accessing anyone else’s data or performing unauthorised attacks.

This exercise develops practical threat-modelling habits while respecting privacy and legal boundaries.

4 November: why essential services need stronger defences

The university also listed a 4 November public broadcast featuring Associate Professor Daisuke Mashima discussing risks to infrastructure, essential services and personal information.

Services involving water, energy, transport or healthcare may depend on specialised operational technology and interconnected IT systems.

An organisation needs to protect both the information it stores and the real-world functions its systems control.

A failure in an ordinary office spreadsheet has different consequences from a failure in a safety-critical industrial control system.

Operational technology versus office computing

Office IT often focuses on documents, communication, databases and business applications.

Operational technology (OT) interacts with physical processes, such as pumps, sensors, industrial controls or other equipment.

Security teams need to understand the devices’ constraints. A routine software update suitable for a laptop may be difficult or risky on a system that must operate continuously.

A reliable security plan coordinates appropriate technical controls with operational safety.

Why incident response matters

A team should know whom to contact, how to preserve evidence, how to restore services and how to avoid amplifying the damage.

A backup is useful only if it is reliable and can actually be restored under the relevant conditions.

Security documentation should be tested through authorised exercises rather than assumed effective because a policy exists on paper.

This makes the connection between Computing, Systems Engineering and real public services clear.

28 October: Sunday wins S$16,000 for assistive AI

The SUTD ISTD achievements page describes a university-linked team led by Rajaselvam Kalyanaraman, with collaborators Kunalkumar Senthilkumar, Pravin R and Devadharshini Kumar, winning Huawei’s Tech4City top award.

The team’s project, Sunday, is an AI assistant intended to help people with visual impairments operate smartphone functions using voice commands.

The reported prize was S$16,000.

The team developed a solution aimed at a real accessibility challenge: many smartphone tasks require navigating visual menus that may be difficult to use with ordinary interfaces.

How might a voice assistant help?

A well-designed assistant could interpret a user’s spoken request, identify the relevant device function and provide useful feedback.

For example, someone might want to find an appointment reminder, open a particular application or change a setting.

The interface must work reliably with screen readers, accessibility permissions and the user’s chosen method of confirmation.

A system that performs a task without explaining what it will change can undermine user control.

Why accessibility needs more than a demonstration

People with visual impairments have diverse needs and preferences.

Some may use existing screen-reader gestures efficiently and find a new AI layer unnecessary. Others may benefit from conversational control for certain tasks.

A tool should therefore be evaluated with real users, with appropriate consent and attention to error recovery.

Prize recognition is encouraging, but winning a competition does not establish that Sunday is already certified, universally accessible or commercially deployed.

The importance of asking permission

A smartphone assistant may access messages, contacts or other private information.

It should request only necessary permissions and explain consequential actions before performing them.

If a user asks the system to send a message, the tool should provide a meaningful opportunity to confirm its recipient and content.

Designing accessible AI includes protecting autonomy and privacy rather than assuming that automation is always helpful.

A school-level app-design exercise

Imagine a student designing a simple reminder application for someone who finds small text difficult to read.

The student can test large controls, clear audio feedback and a way to undo mistakes.

It is important to involve appropriate users rather than make assumptions solely from what the developer finds convenient.

A good project measures whether people can complete the actual task, not just whether the interface looks attractive.

28 September: Bloom Up and practical AI for daily tasks

The ISTD September 2026 news archive describes another socially focused student project, Bloom Up.

SUTD students won second prize in the university category at Dell InnovateFest 2026 for an application intended to help people involved in a Movement for the Intellectually Disabled of Singapore (MINDS) programme carry out everyday tasks more independently.

The result highlights a related design issue: technology needs to respect the abilities, preferences and routines of its intended users.

A programme that appears simple to a developer may be confusing or demanding to someone in a different context.

What makes a useful assistive application?

A well-designed application should have a clearly defined purpose, understandable interaction and a way to recover from mistakes.

It should also protect users’ personal information and allow appropriate assistance when the user wants or needs it.

Success may involve independence, comfort or task completion rather than the number of downloads.

A competition award recognises the prototype and team under the contest’s rules, but does not prove long-term outcomes for every potential user.

6 October: self-driving perception in rain and darkness

The SUTD technical release dated 6 October 2026 describes research led by Assistant Professor Zhao Na.

Autonomous vehicles often use cameras and LiDAR together. Cameras capture appearance and colour, while LiDAR measures distances using laser light.

The combination can be powerful, but the reliability of each source can change under rain, darkness or a different urban environment.

Models trained mainly on clear daytime data may learn to depend too heavily on one sensor, then fail when the conditions change.

The research examined methods for rebalancing and adapting how the two kinds of information are used.

What is complementary collaborative fusion (CCF)?

The CCF method gives the camera and LiDAR branches stronger individual training signals, rather than allowing one branch to dominate.

The researchers also used selected masking during training so the model needed to learn from whichever evidence remained available.

On the nuScenes benchmark dataset, the team reported average-precision improvements of 2.8 percentage points in rainy conditions, 1.3 points at night and 3.2 points on Boston-area tests, relative to the named comparison.

Those are particular benchmark improvements, not proof that a car equipped with the system never makes mistakes.

What is PanDA?

A separate method, PanDA, investigates adaptation of multimodal scene segmentation to conditions where labelled training data are limited.

The system uses provisional labels and two kinds of correction: one drawing on three-dimensional geometry, and another using image-related recognition.

The reported gains in panoptic quality were 13.2 percentage points when transferring Boston-to-Singapore, 8.9 points from sunny-to-rainy and 8.4 points from day-to-night in the evaluated benchmark.

Those results are promising but depend on the dataset and reference systems.

A model can perform well in an evaluation while still requiring more testing in rare real-world traffic situations.

Does this make self-driving cars safe in all weather?

No. Assistant Professor Zhao explicitly said further testing is required across rare events, other cities and different sensor configurations.

Researchers also need to assess processing speed, reliability and safety mechanisms before deployment in production vehicles.

An autonomous-driving model is not a substitute for the applicable regulatory and operational safety framework.

Students should learn to state the difference between research accuracy and certified safety.

A simple scientific analogy

Imagine two students observing an event. One can hear it clearly but has a poor view; the other can see it well but cannot hear all the words.

Combining their observations can improve understanding, provided neither source is treated as infallible.

If conditions change, the value of each observation changes too.

The real technical problem involves complex geometry, machine learning and statistical evaluation, but the analogy explains why complementary information is useful.

24 September: radar learns to estimate movement

The SUTD ISTD research archive describes IterFlow, a framework led by Assistant Professor Zhao Na for interpreting 4D radar observations in traffic scenes.

Radar can help sense distance and motion in conditions where cameras or LiDAR face difficulties.

The research examines how a lightweight model can estimate three-dimensional scene motion from radar measurements.

This is useful because autonomous systems need to recognise not only where an object is, but how it is moving.

What is scene flow?

Scene flow describes how points in a three-dimensional scene move over time.

For a driving system, the motion of a pedestrian, cyclist or vehicle may be more important than an isolated image of where the object was one moment earlier.

Reliable estimation requires understanding noise, incomplete observations and how different objects can move independently.

A method that works on one road dataset may require further validation under different weather and environments.

Why radar is not a complete solution

Radar has strengths for motion sensing but may provide less detailed shape information than some optical sensors.

A vehicle might combine several sensing technologies to overcome individual weaknesses.

The design challenge is deciding which signals are trustworthy in the current conditions and what to do when they disagree.

The goal is safer perception through evidence, not a claim that adding one sensor removes all risk.

21 September: theory-guided AI reduces IoT congestion

The ISTD news archive also highlights a 21 September 2026 research report on a theoretical and AI-guided method for managing random access in Internet of Things (IoT) networks.

The collaboration between Beihang University and SUTD investigated how many devices can share network capacity more efficiently when they attempt to transmit.

The university’s summary says the method halved the time needed to clear congested queues in its evaluated conditions while also improving relevant latency and collision measures.

Why do IoT devices create congestion?

Imagine thousands of small sensors attempting to send information at nearly the same time.

If too many devices transmit using the same shared resource, collisions or repeated attempts may delay delivery.

A system needs to manage access so that important information can move reliably without wasting energy.

This matters for connected industrial equipment, environmental sensors and other systems operating at scale.

What is theory-guided AI?

A purely data-driven model may discover a useful pattern while being difficult to interpret.

Theory-guided AI incorporates mathematical knowledge about the system’s structure, which can improve efficiency and help constrain decision-making.

For network access, mathematical analysis can guide how the system responds to queues, traffic and transmission conditions.

That does not automatically make an algorithm optimal for every possible network. The assumptions and deployment environment matter.

A simple numerical example

Suppose a network clears a particular queue of sensor messages in 20 seconds under one baseline.

If an improved method halved that duration under the same test, the time would be 10 seconds.

This is a hypothetical illustration of what “halved” means, not the exact measured data from the SUTD study.

A careful student should check whether the comparison uses the same traffic level, hardware and conditions before interpreting the improvement.

10 September: HieraScaffold and realistic LiDAR simulation

The ISTD news and research highlights describe HieraScaffold, an AI framework for generating more coherent 4D LiDAR scenes.

The work aims to make it easier for autonomous driving and robotics teams to develop dynamic simulation data.

Real-world data collection can be expensive and may not cover rare situations. Simulated scenes could help evaluate algorithms under more varied conditions.

But a simulated environment remains a model. Engineers need to test whether its statistics and physical behaviour match the kinds of situations that matter in practice.

How simulations can mislead

A simulation might create visually convincing pedestrians and vehicles while failing to represent the actual timing of movement or the behaviour of sensors in rain.

A model trained on such scenes could perform well inside the simulation and less reliably on actual roads.

Researchers therefore need to compare simulated data with real measurements, examine gaps and report where generalisation remains uncertain.

Simulation is most useful when it supports careful verification, not when it allows a team to skip every real-world test.

What is the ISTD undergraduate degree actually called?

The SUTD ISTD curriculum page states that the current qualification awarded through the pillar is Bachelor of Engineering (Computer Science and Design), commonly abbreviated CSD.

It explicitly explains that the degree was previously named Bachelor of Engineering (Information Systems Technology and Design).

This distinction is crucial for a 2027 applicant. The academic school or pillar continues to use ISTD, but the student should select the current CSD undergraduate degree in official programme information.

The current SUTD majors overview lists five main undergraduate degrees: Architecture and Sustainable Design, Computer Science and Design, Design and Artificial Intelligence, Engineering Product Development, and Engineering Systems and Design.

Computer Science and Design versus Design and AI

CSD develops deep computing foundations, with flexibility across software, AI, data, security and related systems.

Design and Artificial Intelligence (DAI) connects AI with design and a wider range of questions about people, organisations, business and society.

They overlap in AI learning but are not identical in compulsory technical content.

A student interested in becoming a software engineer should examine CSD’s algorithms and systems requirements; someone interested in AI-driven services or human-centred design should compare DAI’s applied curriculum.

Choosing by the word “AI” alone is insufficient.

The eight-term SUTD undergraduate structure

The SUTD undergraduate overview describes an eight-term programme beginning in September.

Terms 1–3 form the shared Freshmore foundation, combining design, Mathematics, programming, Physics and the Humanities.

Students then progress into their selected major, develop technical depth and undertake interdisciplinary projects.

The university describes a grade-free first term, along with options to take specified additional subjects on a grade-free basis in Terms 2 and 3 under the current rules.

Why the common curriculum matters

A Computing student may later need to build software for hospitals, transport systems or public services.

Those applications involve people, safety, legal rules and institutional constraints as well as algorithms.

An interdisciplinary foundation helps students ask about the full problem rather than assume that a technically working program is automatically an appropriate solution.

It also gives students time to explore which major best matches their interests.

CSD curriculum: core, electives, Humanities and capstone

The ISTD curriculum description identifies CSD core subjects, flexible technical electives, Humanities and Social Sciences requirements and a final-year capstone.

The published framework describes 60 credits in CSD Core, 96 technical elective credits, 60 HASS credits and 24 capstone credits within the relevant degree structure.

It notes that the CSD core is concentrated in Terms 4 and 5, while electives allow customisation in later study.

Students entering in AY2027 must verify whether any requirements have been revised from the published structure for their particular cohort.

What does the CSD core develop?

The core provides foundations in computational thinking, Algorithms, Mathematics, Data and Systems.

Students need to know how programs represent information, how computation scales with problem size and how to test whether a system behaves correctly.

AI can help write code, but it cannot replace understanding why the code is safe or efficient.

Technical depth becomes especially important when software affects personal information or physical infrastructure.

Why Humanities belongs in a Computing degree

A student designing a digital identification service needs to understand privacy, access and public trust.

A student building an AI education tool should consider how the system changes learning, rather than measure only speed.

Humanities and Social Science courses help students think about ethics, institutions and human behaviour.

They do not replace technical learning; they help define why technology is being built and which consequences matter.

CSD specialisation tracks

The ISTD undergraduate track list includes these named directions:

• Artificial Intelligence.
• Data Analytics.
• Financial Technology.
• Internet of Things and Intelligent Systems.
• Security.
• Software Engineering.
• Visual Analytics and Computing.
• Custom Track under approved academic planning.

A track guides technical elective choices and can be subject to conditions. It is not automatically a second bachelor’s degree.

Students should check the actual course prerequisites and the availability of modules before deciding.

Security specialisation

A Security student may investigate cryptography, secure software, networks and system vulnerabilities.

The aim is to build and defend authorised systems, not gain permission to penetrate other people’s accounts.

The 10 November cybersecurity report shows why technical ability must be combined with organisational policy and ethical judgement.

Software Engineering specialisation

Software Engineering focuses on designing, testing, deploying and maintaining software systems.

A successful application must handle real users, changing inputs, errors and future maintenance.

The field is not simply creating a screen that looks polished or producing code rapidly with AI.

Artificial Intelligence specialisation

AI study may involve machine learning, optimisation, representation and evaluation.

A good student learns how models are trained, why they fail and how to assess claims of accuracy.

The autonomous-driving research demonstrates that performance can change under new weather, location or sensor conditions.

Data Analytics specialisation

Data Analytics involves processing, modelling and interpreting data.

A student should be able to distinguish missing observations, biased samples and causal claims that the data do not establish.

The best analysis provides an understandable explanation of uncertainty.

Financial Technology specialisation

Fintech combines computing with systems involving financial transactions, banking operations or related services.

Security, compliance and reliability are important because errors can affect real funds or customer rights.

A Fintech track is not an automatic licence to provide regulated financial advice.

IoT and Intelligent Systems

IoT systems link sensors, devices and networks.

The September congestion research illustrates why devices need efficient, secure communication rather than assuming unlimited wireless capacity.

Engineers should also consider how to update devices and handle failure after deployment.

Visual Analytics and Computing

Visual Computing can involve computer vision, graphics, spatial reasoning and analysis of image or sensor information.

The 2026 driving research demonstrates how visually based algorithms can support real-world tasks while remaining sensitive to environmental changes.

A career in this area may require advanced Mathematics and substantial programming skills.

Applying for SUTD Computer Science and Design in 2027

The SUTD undergraduate admissions page reviewed on 12 November 2026 still displays the 2026 admissions exercise.

For reference only, it lists a 19 March 2026 A-Level deadline and 2 January–2 March 2026 for several other qualification groups.

Those dates have passed. It would be misleading to present the same day and month with 2027 added unless the university confirms them.

Applicants should monitor the official qualification-specific application timeline and the correct portal.

Which applicants can be considered?

SUTD accepts suitable applicants with recognised Singapore-Cambridge A-Levels, local polytechnic diplomas, IB, NUS High School and other international qualifications.

The university states it generally prefers good foundations in Mathematics and Sciences, particularly Physics or Chemistry, while reviewing applications holistically.

Students should also be prepared to demonstrate technical interest through truthful projects or academic work.

Some applicants may be invited for a conversation with faculty, depending on the admissions process.

Preparation checklist

1. Identify the correct qualification and SUTD application route.
2. Check the current BEng Computer Science and Design title.
3. Review the Freshmore sequence and CSD core.
4. Compare CSD with Design and Artificial Intelligence.
5. Prepare actual academic transcripts and any permitted supporting projects.
6. Monitor the real 2027 undergraduate deadline.
7. Investigate scholarships, fees, housing and practical learning.
8. Submit through SUTD’s official application portal and keep confirmation.

No paid service can guarantee acceptance or bypass formal requirements.

Tuition and financial-aid planning

SUTD’s official tuition fee page states that eligible subsidised Singapore citizens entering in AY2026 pay S$13,500 annually, under the university’s cohort-based fee model.

The same table lists subsidised rates for eligible Singapore permanent residents and international students, with separate non-subsidised amounts.

An eight-term degree normally spans four academic years, so four years at the stated Singapore citizen rate would equal S$54,000 in tuition, before compulsory miscellaneous fees and living costs.

That amount is a historical cohort reference. It is not a binding AY2027 tuition quote.

Miscellaneous costs

SUTD’s AY2026 schedule lists compulsory miscellaneous fees of S$165 per term for Singapore citizens and permanent residents.

Students should also budget for transport, meals, accommodation if required, computing devices, approved technical activities and overseas opportunities.

A scholarship or bursary reduces costs only under its specific written conditions. A student loan is normally repayable.

For the wider financing context, see eduKate’s Singapore university aid guide.

CSD and careers in an AI-driven economy

Graduates may explore Software Engineering, security, data, AI, financial technology, systems development, product roles and research, depending on projects and technical depth.

Not every technology job is identical. A network security engineer needs skills different from a machine-learning researcher or user-interface developer.

The SUTD ISTD news archive also discusses the challenge some fresh graduates face in finding jobs that match their expectations and available experience.

A degree, research ranking or contest award does not guarantee employment. Practical projects, ability to explain code and willingness to learn remain important.

The value of independent problem solving

AI can create a draft program quickly, but a graduate needs to know whether it handles failures, respects permissions and produces the correct result.

A candidate who explains how a project was tested may demonstrate stronger professional readiness than someone presenting a larger system they cannot understand.

This is particularly important when AI tools make it easy to produce convincing but insecure code.

Full Subject-Based Banding and the pathway to Computing

Full Subject-Based Banding concerns Singapore secondary subject levels G1, G2 and G3. These are not direct SUTD Computer Science and Design admission categories.

A student later needs an accepted post-secondary qualification and relevant academic preparation.

Mathematics, Science, English and Programming can develop foundations.

A secondary learner can practise by building a safe offline program, writing tests and explaining what happened when it failed, rather than attempting unauthorised cybersecurity experiments.

A PSLE score does not permanently determine future university eligibility.

Three reader decisions

A JC student interested in security: investigate CSD’s Security track, Mathematics requirements and the difference between defending systems and performing authorised assessments. Verify the eventual 2027 admissions window.

A polytechnic student interested in AI and robotics: compare CSD with Engineering Product Development and DAI. Decide whether you prefer deep software foundations, cyber-physical product design or broader human-centred AI.

A parent worried about cybercrime and AI: the CSA figures show why basic security matters, but not all incidents were AI-related. Encourage digital safety, critical reasoning and appropriate permissions, not fear of every new technology.

Frequently asked questions

Is ISTD still a school at SUTD?

Information Systems Technology and Design remains the name of the academic pillar. Its undergraduate degree is now BEng Computer Science and Design.

What was the key November 2026 cybersecurity finding?

A CSA report highlighted by SUTD says 55% of 2,005 surveyed organisations used AI while the surveyed organisations implemented 39% of relevant basic AI security measures on average.

Did 90% of organisations experience AI-specific cyberattacks?

No. The report says nearly 90% experienced at least one cybersecurity incident; not all incidents were AI-related.

What is Shadow AI?

Employees’ use of AI systems without appropriate organisational approval, awareness or controls.

What is Sunday?

An AI-assisted smartphone accessibility project whose university-linked team won a S$16,000 Tech4City prize in October 2026.

Does the award mean Sunday is a commercially validated medical device?

No. Prize recognition does not establish universal deployment, regulatory approval or clinical outcomes.

What did SUTD improve for self-driving cars in 2026?

Researchers developed methods for more robust camera–LiDAR sensor fusion and adaptation in rain, darkness and different cities, under specific benchmark conditions.

Is the current computing degree eight terms?

Yes. SUTD’s usual undergraduate programme spans eight terms, beginning with a common Freshmore foundation.

Which specialisation tracks are listed?

Artificial Intelligence, Data Analytics, Financial Technology, IoT and Intelligent Systems, Security, Software Engineering, Visual Analytics and Computing, and an approved Custom Track.

What is the AY2027/28 undergraduate application deadline?

The official SUTD pages checked still showed the 2026 admissions cycle, so the actual 2027 date must be confirmed from the university when released.

How much is undergraduate tuition for an eligible Singapore citizen?

The published AY2026 subsidised rate is S$13,500 per year, subject to its cohort rules. The AY2027/28 rate needs separate verification.

Can Full SBB students eventually study CSD?

Potentially through an eligible later qualification and current university admission requirements. Secondary levels alone do not determine admission.

Official sources and next reading

• SUTD ISTD academic pillar and current news
• 10 November 2026 AI adoption and cybersecurity report
• SUTD ISTD 2026 news archive including November and September
• SUTD ISTD achievements, including Sunday Tech4City prize
• SUTD autonomous-driving sensor research, 6 October 2026
• SUTD 2026 research and LiDAR developments
• SUTD CSD undergraduate curriculum and degree renaming
• SUTD CSD technical and specialisation tracks
• SUTD five undergraduate majors
• SUTD AY2026-onward Freshmore and capstone structure
• SUTD undergraduate admissions page, pending 2027 update
• SUTD qualification-specific application timeline
• SUTD AY2026 entering-cohort tuition and MOE grants
• eduKate: SUTD Engineering Product Development 2026 news
• eduKate: SIT Infocomm Technology 2026 news and 2027 admissions
• eduKate: Singapore university financial aid

Editorial check, 12 November 2026: The CSA survey’s denominators and incident qualifications are preserved, research benchmark gains are not represented as zero accident risk, and the university’s CSD degree is not misnamed after its ISTD pillar. The current undergraduate site still advertises 2026 admissions dates, so no invented AY2027 closing date appears. Award recognition is not equated with clinical approval, commercial deployment or guaranteed employment.