How to Write the IY2S510 AI Policy and Risk Matrix (Assessment 2, USW)

Last updated on: November 19, 2026 | Students Assignment Help UK Editorial Team | 3 views

A strong IY2S510 Assessment 2 answer reads like a real organisational policy, not an essay. It sets out a short list of principles, names the risks PoliceAssist creates at Cooper’s Chase Police Force, ties each risk to a mitigation and an owner, and finishes with a risk matrix whose scores make sense. The risk matrix and the risks with their mitigations carry 65% of the marks between them, so build those first and write the policy wording around them.

If you would like a second opinion on your structure or scoring, our risk management assignment help page explains the support available.

What the assessment is really asking

The IY2S510 Compliance and Risk Management Assessment 2 brief at the University of South Wales gives you an information pack on a fictional police force that wants an AI assistant to help with non-emergency online reports. PoliceAssist would ask follow-up questions, write a structured summary and suggest a category and priority. A trained member of staff would review each suggestion before anything happens.

Your job is to write the policy that would let the force introduce that system safely. The brief asks for four things: the current risks, the mitigations, how the system fits into the force’s systems and staff with clear accountability, and a risk matrix. It is a 1,500-word individual project worth 70% of the module, referenced in USW Harvard.

Marks and words at a glance

Criterion Weighting Suggested words
Risk matrix 35% A table plus about 150 words
Risk identification and mitigations 30% About 550
Accountability and integration 15% About 350
Policy language and formatting 15% Applied throughout
USW Harvard references 5% Not usually counted, but check your brief

The suggested split simply follows the weightings. Purpose, scope and principles should take no more than about 300 words, because they earn marks mainly through the language and formatting criterion.

Want to see the shape of a finished answer first? Our IY2S510 PoliceAssist policy and risk matrix sample is for information and reference only, and every score in it is labelled illustrative.

Write it as a policy, not an essay

Fifteen per cent of the marks go to policy language and format. Markers want something a staff member could follow. That means numbered sections, short clauses, plain words and a consistent voice such as “The force will” or “Reviewers must”. Long paragraphs of discussion, rhetorical questions and first-person opinions belong in an essay, not here.

A clear outline might be: purpose and scope, guiding principles, risks and mitigations, risk matrix, integration and accountability, monitoring and review. Keep the scope tight. PoliceAssist handles non-emergency online reports, so say plainly that it does not replace 999 and does not make final decisions.

Build principles you can actually use

The brief asks you to research existing AI policies, so your principles should show that you have done so. Useful UK sources include the government’s cross-sector AI principles, the national policing AI covenant, Information Commissioner’s Office guidance on AI and data protection, and NCSC guidance on secure AI development. You do not need to copy their wording. Instead, translate each principle into a rule for Cooper’s Chase.

For example, “human oversight” becomes “no report is closed or downgraded without a trained reviewer”. “Safety” becomes “any sign of immediate danger triggers a 999 prompt”. Five to seven principles are enough, and a two-column table keeps them short.

Find risks that belong to this scenario

Generic AI risks such as “AI may be biased” earn little on their own. The marks go to risks that come from the information pack. Read it again and list what could go wrong at each point PoliceAssist touches.

Where the risk comes from Example risk Example mitigation
Follow-up questions A sign of immediate danger is missed or worded indirectly Mandatory danger questions, automatic 999 prompt, timed reviewer check
Summaries Wrong location, time or suspect detail in the summary Keep the original text, reviewer comparison, accuracy sampling
Category and priority Bias by area or group, or misplaced trust in the score Bias testing, monitoring, reviewer training on automation bias
Sensitive personal data Breach or misuse at the cloud AI supplier Data protection impact assessment, contract terms, retention limits
Public website Prompt injection, data extraction or flooding Input filtering, rate limiting, penetration testing
Incident management link Supplier route into police records One-way transfer of reviewed reports, least privilege, logging
Staff and public Low trust, deskilling or fear of job loss Transparency notice, human option, training for new reviewer roles

For each risk, say what the mitigation actually changes. Does it lower the likelihood, reduce the impact, or both? That link is what turns a list into analysis.

Get the law right for policing data

Many students mention UK GDPR and stop there. In a police setting, processing for law enforcement purposes falls under Part 3 of the Data Protection Act 2018, while other processing, such as some website analytics, may fall under UK GDPR. Showing that you know the difference is an easy way to stand out. Mention the role of the Data Protection Officer, a data protection impact assessment before launch, data minimisation and retention, and the force’s equality duties.

Keep this practical. One or two sentences per point, tied to PoliceAssist, beat a long summary of the legislation.

Make the risk matrix coherent

The matrix carries 35% of the marks, the largest single share. A coherent matrix has a stated scale, ratings that follow from that scale, and a clear link between each mitigation and the residual score.

  1. State the scale first, for example likelihood 1–5 and impact 1–5, with the rating bands you use.
  2. Score each risk before mitigation, then after mitigation.
  3. Give a short reason for any high score, especially the danger-detection risk.
  4. Assign one named owner per risk, using the roles in the information pack.
  5. Check that the residual rating only drops where the mitigation really affects likelihood or impact.

Avoid making every residual risk “low”. Some risks, such as loss of public trust, are hard to remove completely, and markers notice when the scoring looks too convenient.

Accountability and integration

The information pack gives you the roles, so use them. The Chief Constable carries ultimate responsibility, the Chief Information Officer is the Senior Responsible Owner, the Head of Information Security owns cyber controls, the Data Protection Officer advises on personal data and the Head of Contact Management runs the control room and the reviewer team. A short RACI-style table or a set of numbered clauses works well.

For integration, describe a phased launch: a closed test with synthetic reports, a limited pilot with every output reviewed, then a wider roll-out only after agreed safety and accuracy thresholds are met. Keep a manual reporting route in place throughout. Our IT management assignment support covers this kind of system integration planning.

Mark-maximising checklist

  • Policy format with numbered sections and plain, directive language.
  • Five to seven principles, each turned into a rule for Cooper’s Chase.
  • Six to eight risks drawn from the information pack, not a generic list.
  • Every mitigation linked to a lower likelihood, a lower impact or both.
  • Part 3 of the Data Protection Act 2018 recognised for law enforcement processing.
  • A risk matrix with a stated scale, inherent and residual ratings and owners.
  • Named roles from the information pack for every accountability point.
  • A monitoring and review section with clear review dates.
  • USW Harvard references to current UK law, regulators and AI frameworks.

Common mistakes to avoid

  • Writing a discussion essay about AI in policing instead of a policy.
  • Listing risks without saying how each mitigation changes the score.
  • Scoring every residual risk as low, or using a scale that is never explained.
  • Forgetting the human reviewer, who is central to the force’s design.
  • Treating the supplier’s security as given instead of something to verify.
  • Copying wording from published AI policies rather than adapting it.

Where students usually get stuck

Many students find the 1,500 words tight once the matrix is in. Tables help, because they carry detail without long sentences. Others struggle to decide whether a risk is a likelihood problem or an impact problem, so try writing one sentence for each: “this could happen because…” and “if it happens, the harm is…”. A third issue is AI use. The brief allows AI for learning and planning only, so the policy and matrix must be your own writing.

For help with structure and referencing, our report writing help is a useful next step, and you can browse more University of South Wales assignment briefs.

FAQs

How many risks should the IY2S510 risk matrix include?

Six to eight well-explained risks usually work within 1,500 words. Fewer can look incomplete, while many more leave no room to explain mitigations.

Which risk should be rated highest?

Most answers rate a missed sign of immediate danger highest, because the impact is severe even if the likelihood is moderate. Whatever you choose, justify the score.

Do I need to mention the Data Protection Act 2018?

Yes. PoliceAssist processes victim and suspect data, and law enforcement processing falls under Part 3 of the Act, so it is central to the data protection risks.

Can I use a 3×3 matrix instead of 5×5?

Either can work if you explain the scale and apply it consistently. A 5×5 scale gives more room to show the difference between inherent and residual risk.

Who should own the overall risk register?

The information pack names the Chief Information Officer as Senior Responsible Owner, so that role is the natural owner, with specialist risks assigned to the relevant heads.

Can I use AI tools for this assessment?

The brief sets the work at a level where AI can support learning and planning, but the submitted policy and matrix must be written by you. Follow your module guidance.

If you are short of time before the deadline, send your details through our submit assignment page and we will advise on the next step, or read about us to see how we work.

Written by
Students Assignment Help UK Editorial Team

Students Assignment Help UK Editorial Team

Our editorial team works with students in the UK and sees the questions they ask most often. This guide is built from those real queries: where students get stuck, what markers look for and how to plan the work step by step.

Latest Posts

UPTO 35% DISCOUNT

Get Your Assignment Completed At Lower Prices

Plagiarism Free Solutions
100% Original Work
24*7 Online Assistance
Native PhD Experts
Hire a Writer Now

Facing Issues with Assignments? Talk to Our Experts Now!Download Our App Now!

Have Questions About Our Services?
Download Our App!

Get the App Today!

QRcode