SCHOOL OF LAW

COURSEWORK COVER SHEET (2025-26)

ASSIGNMENT DETAILS
Module Title & Code: LAW3134: Law and Technology
Student Number: 40410963
Assignment’s Full Title: Law3134 Assignment 2

WORD COUNT
(excluding bibliography, but including footnotes)

Assigned word count for the essay: 750

The actual number of words in the essay: 709

Penalties will be imposed for work that exceeds the prescribed word limitation. The penalties for exceeding the word limitation are:
 If a student exceeds the word limit by one word, or any number of words equivalent to an excess of up to 10% more than
the limit, 5 marks will be deducted;

 For each subsequent 10% or part thereof, 5 further marks will be deducted.

Word count penalties will be applied strictly.

SUBMISSION DEADLINE

Deadline: by 12 noon on: 21/11/2025

in line with University and School of Law Regulations, the penalty for late submission is a reduction of 5 marks for each
calendar day late or part thereof, up to a maximum of 5 calendar days late. Work received after that point will receive a mark
of 0.

UNIVERSITY REGULATIONS
By submitting the work, I declare that:
1. I am fit to be examined
2. I have read and understood the University regulations relating to academic offences, including collusion and plagiarism:
http://www.qub.ac.uk/directorates/AcademicStudentAffairs/AcademicAffairs/GeneralRegulations/Procedures/
ProceduresforDealingwithAcademicOffences/
3. The submission is my own original work and no part of it has been submitted for any other assignments, except as otherwise
permitted;
4. All sources used, published or unpublished, have been acknowledged;
5. I give my consent for the work to be scanned using a plagiarism detection software.
 MEMORANDUM
TO: The Belfast Bot
FROM: Henry O’Riordan
DATE: 21/11/2025
RE: The Belfast Bot’s AI Legal Advice App

Generative AI Legal Advice

Regarding the first issue, the application, in its current form, is permissible as legal advice is
not a “reserved activity” i. However, should the company make itself out to be competent and
declare that generative AI output “is all people need to understand law”, it will then owe
clients a duty of care, thus it could find itself liable to pay damages in potential instances of
negligence.

Section 49 of the Consumer Rights Act 2015 iirequires that all services are provided with
“reasonable care and skill”. Wright v Troy Lucas and Co iiiestablishes that if an unregulated
legal service provider gives sufficient reason/advertises as a competent adviser, they inherit a
duty of care and thus a liability for any damages that a client incurs because of negligence. In
Ayinde v Haringey LBCiv, Zzaman v Revenue and Customs Commissionersv, and Harber v
Revenue and Customs Commissionersvi, reliance on inaccurate, AI-generated citations led to
the dismissal of appeals in the instances of Ayinde, Zzaman, and Harber.

The Belfast Bot’s AI model for legal advice has an increased chance of producing inaccurate
and harmful advice. Generative AI provides legal analysis based on training data and
probability and does not operate as a search engine or authoritative databasevii. Additionally,
an absence of qualified oversight disproportionately increases risks for unregulated service
providers using AIviii. Whilst there is a growing demand for services like the Belfast Bot, and
certainly some positives to the accessibility and efficiency of unqualified legal advice, the
capacity to identify inaccuracies in data is significantly reduced without qualified oversight. ix

The company’s intention to “eliminate the need to consult expensive legal professionals”
maintains a high level of risk for the operation of its services. I recommend the
implementation of robust accountability and governance of AI systems overseen by a
qualified legal professional to mitigate risks, as recommended by governing legal authorities
and jurisprudence in the area. x

Data Privacy

On the second issue, the collection and usage of personal information is permissible and risk
averse, provided The Belfast Bot fulfils its statutory obligations relating to consent,
transparency, and processing. The subcategories include the proper procedure for obtaining
user consent and ensuring that the use and storage of data are compliant with the relevant
statutes.

Article 4(1) of the UK GDPR xioutlines that consent of data subjects must be freely given,
informed, specific, and affirmative. Additionally, UK GDPR requires a Data Protection
Impact Assessment for high risk data xii. Article 5(1)(a) of the Data Protection Act 2018 xiii
lawfully and transparently. Harrison v Cameron & Alasdair Cameron Ltd xivand Ashley v

2
HMRC xvhighlighted the right of data subjects to know how their data is being communicated
and to whom.

In the context of The Belfast Bot, the input of personal user information into AI systems is
the first key risk. Third party access is often characteristic of Generative AI systems, meaning
personal information of clients may be at risk. Secondly, storage of data on “servers across
the globe” may further risk of data leaks or unexpected third party access.

Firstly, amending the consent procedure to inform users that data will be accessed by third
parties is a key step to ensuring compliance and mitigating liability. Secondly, conducting
regular Data Protection Impact Assessments to guarantee client data is shared appropriately.

3
i
Legal Services Act 2007, s 12,
ii
Consumer Rights Act 2015, s 49,
iii
Wright v Troy Lucas (a firm) & Rusz EWHC 1098 (QB),
iv
R (Ayinde) v London Borough of Haringey EWHC 1040 (Admin),
v
Zzaman v Commissioners for His Majesty's Revenue and Customs UKFTT 00539 (TC),
vi
Harber v Commissioners for His Majesty’s Revenue and Customs UKFTT 1007 (TC),
vii
Michael Legg and Vicki McNamara, ‘Generative AI, Fake Law and Professional Guidance’ (2024) UNSW
Law Research No 24-34 https://ssrn.com/abstract=4592622 accessed 15 November 2025,
viii
IRN Legal Reports, ‘Unregulated Legal Providers Literature Review for CILEx Regulation Ltd’ (2024)
https://cilexregulation.org.uk/wp-content/uploads/2023/04/IRN-Legal-Reports-Unregulated-Legal-Providers-
CILEX-Literature-Review- accessed 15 November 2025,
ix
Id,
x
Solicitors Regulation Authority, ‘Risk Outlook report: The use of artificial intelligence in the legal market’ (20
November 2023) https://www.sra.org.uk/sra/research-publications/artificial-intelligence-legal-market/
accessed 15 November 2025,
xi
UK General Data Protection Regulation (UK GDPR) art 4(1),
xii
Id,
xiii
Data Protection Act 2018 s 5(1)(a),
xiv
Harrison v Cameron & Alasdair Cameron Ltd EWHC 1377 (KB),
xv
Ashley v HM Revenue and Customs EWHC 134 (KB),