AI Use Case Intake & Approval Procedure

AI governance lead

[[e.g. Head of Risk]]

Owns the AI policy and the approved list, screens every proposed use, keeps the inventory, approves Tier 1 uses and Tier 2 uses (with the Head of Information Security), may suspend an approved running use pending review, advises on Tier 3 and reports the measures.

AI use owner

[[the manager accountable for the business activity the AI supports]]

Proposes the use, answers the intake questions, answers for its output, names the people who check it, is consulted on the decision and signs the conditions it accepts, and asks for a review when anything changes. Never approves a use it owns.

Head of Information Security

[[e.g. Head of Information Security]]

Assesses the security factor, agrees the tool's conditions and approves Tier 2 uses with the AI governance lead.

Data Protection Officer

[[or the privacy lead where no DPO is appointed]]

Advises on personal data in any use, decides with the use owner whether a data protection impact assessment is needed, and is consulted on every use of personal data.

Legal

[[in-house or external counsel]]

Confirms the AI Act role and category of each system, reads supplier terms, and decides whether the organisation could be a provider.

HR

[[e.g. HR Business Partner]]

Advises on any use that touches recruitment, pay, performance or monitoring of staff, and on telling and consulting staff or their representatives.

Procurement

[[e.g. Procurement Manager]]

Makes sure no AI tool or AI feature is bought or switched on without an intake decision, and puts the AI terms into the contract (ST-1 to ST-8).

System owner

[[the manager accountable for the system]]

Owns the system an AI feature sits in, and tells the AI governance lead when a supplier adds or changes an AI feature.

Head of IT

[[e.g. Head of IT]]

Connects approved tools to single sign-on, applies the tool's settings and keeps the technical list of what is installed.

Executive decision-maker

[[e.g. Executive Committee, or the management body]]

Decides Tier 3 uses: executive management by default, or the board where the governance structure puts it. Signs as well for a High risk outside appetite (P05). Does not accept a Critical risk as executive management: it goes to the board or its equivalent (P05 RM-04).

Line manager

[[each person's manager]]

Gives time for the literacy plan, checks that the team uses only approved tools, and sends proposals to the use owner.

Awareness lead

[[e.g. Information Security Manager]]

Runs the literacy modules and keeps the completion record, with the learning system.

User

[[everyone who uses an AI tool for work: staff, contractors and supplier staff]]

Uses only approved tools within their data limits, checks the output, discloses AI use where the policy says, and reports mistakes.

IS-1

Propose

The AI use owner fills in the intake questions (IQ-01 to IQ-12) for the tool, the feature or the use, before buying, switching on or trying it with real data.

AI use owner

AG-09

IS-2

Screen

The AI governance lead screens the answers within [[5]] working days of a complete request: prohibited or special category data (Stop), provider signals, then the tier.

AI governance lead

AG-09

IS-3

Assess

At Tier 2 and Tier 3, assess the six factors on the risk and impact assessment (A06). Get the supplier's decision under the third-party pack where Internal or Confidential data is involved, or the use is Tier 2 or 3. The Data Protection Officer advises where personal data is used.

Head of Information Security

AG-11

IS-4

Decide

The approvers of the tier decide: Approve, Approve with conditions, Hold or Decline. Any approver may hold or decline. Only the tier's approvers may approve, and nobody approves a use they own: the use owner is consulted and signs the conditions it accepts. Where a tier's approver owns the use (the AI governance lead, or the Head of Information Security), the executive decision-maker takes that approver's place. Tier 3 goes to the executive decision-maker the procedure names. If the executive decision-maker owns a Tier 3 use, the board or its equivalent decides it (D-42).

The approvers of the tier

AG-09

IS-5

Record

Enter the use in the inventory (A05) with its tier, approver, conditions and review date. Update the approved list (A07) and the supplier register if a supplier is involved.

AI governance lead

AG-10

IS-6

Enable

Connect the tool to single sign-on, apply its settings, and give access only to people who have done the AI essentials module (A09, AL-01). Procurement adds the AI terms ST-1 to ST-8 to the contract.

Head of IT

AG-18

IS-7

Review

Review at the tier's interval, and within [[10]] working days of a trigger: a new purpose, data or audience, a supplier change, an incident, or a change in law. An unapproved use found is stopped within [[3]] working days.

AI use owner

AG-10

IQ-01

What is the use, in one sentence, and what does the business gain?

IQ-02

Which tool or feature, from which supplier (the P06 reference, or none)? Is it already running? What do the supplier's assessment, contract and processor terms say?

IQ-03

Which manager is the AI use owner, and who reads the output before it is used?

IQ-04

What is the highest level of information that goes in (P15)?

IQ-05

Could special category personal data go in (health, trade union membership, biometric data and the others of GDPR Article 9)?

IQ-06

Could any personal data go in or come out?

IQ-07

Do customers or the public deal with the system, or see its output directly, or does a person review anything that leaves the organisation?

IQ-08

Does a person read every output before it is used, or does the system act or send output on its own?

IQ-09

Does the use inform a decision about a person, or decide or shape one (recruitment, pay, performance, discipline, credit, access to a service)? A use that decides or shapes one is Tier 3; a use that only informs one is Tier 2.

IQ-10

If the output were wrong or the tool stopped, would a core service or a material sum be affected (more than P05's Minor level in any impact area)?

IQ-11

Does the use match, or might it match, any practice on the prohibited-practice checklist (A08, PP-01 to PP-10), or fall in an area of Annex III (A08, HA-1 to HA-8)?

IQ-12

Would the organisation build an application or feature on a supplier's model API (prompts, retrieval, tools or code around a model it calls), whether for staff or for customers, fine-tune for others, brand, sell or supply the system, or change what it is meant for? Using a supplier's finished product as supplied is not building. Is it already running?

S-1

Check the request is complete and the owner is a manager who answers for the use.

Return it if not.

S-2

Check for a prohibited practice and for special category data.

Either one is a Stop. Do not go further. Tell Legal for a prohibited practice, and tell the Data Protection Officer for special category data.

S-3

Check for a provider signal (a build on a supplier's model API counts) and for an Annex III area.

Tell Legal. A new use is Tier 3 and held until Legal records which role the organisation holds. If the role is deployer, the tier is then set again from the answers. If it is provider or other, the hold clears and the use stays Tier 3. A use already running carries H-4 as a condition with a date.

S-4

Set the tier from the answers, taking the highest tier any trigger gives. Record every trigger that applied, with its rule, not only the highest.

The tier is set.

S-5

List the holds that apply to a new use (H-1 to H-4), or the dated conditions that take their place for a use already running.

A hold stops the start. A dated condition does not.

S-6

List the assessments, the baseline conditions and the approvers the tier needs, and tell the use owner what happens next and when.

The screen is recorded.

Stop

The request ticks any prohibited practice (PP-01 to PP-10).

Stop

Special category personal data could go in.

Tier 3

The request falls in an Annex III area (HA-1 to HA-8).

Tier 3

The use decides or shapes a decision about a person.

Tier 3

The system acts, or sends output, without a person reading it first.

Tier 3

The organisation would build an application or feature on a supplier's model API, brand, supply or change the purpose of the system, the use is new, and Legal has not yet recorded the role.

Tier 3

The same, but Legal has recorded the provider role or another role: H-4 clears as a hold, and the use stays Tier 3 (the deployer role sets the tier again from the answers).

Tier 2

Confidential information could go in.

Tier 2

Any personal data could go in or come out.

Tier 2

Customers or the public deal with the system or see its output.

Tier 2

The use informs a decision about a person.

Tier 2

A wrong output or a stoppage would affect a core service or a material sum.

Tier 1

None of the above: Public or Internal information with no personal data, staff only, a person reads every output, no effect on a person, and a low operational effect.

Tier 1

A screen against the intake questions. No assessment.

AI governance lead

[[5]] working days

[[12]] months

Tier 2

The risk and impact assessment (A06), the supplier decision under P06 where a supplier is involved, and the Data Protection Officer's advice where personal data is used.

AI governance lead and Head of Information Security, with the use owner consulted (the owner signs the conditions it accepts and does not approve) and the Data Protection Officer's advice recorded where personal data is used

[[15]] working days

[[6]] months

Tier 3

Everything in Tier 2, the EU AI Act worksheet (A08) confirmed by Legal, a data protection impact assessment where Article 35 of the GDPR applies, and HR where staff are affected.

The executive decision-maker your procedure names (executive management by default; the board or its equivalent where your governance structure puts it), on the AI governance lead's recommendation

[[30]] working days

[[6]] months

Approve

Approved

The tier, the approvers, the decision date and the review date. The baseline conditions apply and nothing further.

Approve with conditions

Approved with conditions

As above, with each further condition written so that someone can check it. A condition with a date is a condition of the approval: the tool is switched off if it is not met.

Hold

Held

The reason, the hold that is open (H-1 to H-4) and the date it is due to clear. The use does not start. When the hold clears, the approvers decide again.

Decline

Declined

The reason, in words the requester can act on. The requester may propose again when the reason no longer applies.

H-1

The use has a supplier with no documented decision under the third-party pack yet, and either Internal or Confidential data would go in or the use is Tier 2 or 3 even with Public data (P06 TP-01, TP-05; P15 CL-16 asks for it for Confidential data only, and this procedure asks earlier). A use with no supplier raises no H-1: there is nothing to assess.

The supplier's assessment ends in Approve or Approve with conditions.

H-2

Internal or Confidential data would go in, and the contract does not carry a term barring the supplier from training on it (ST-1; P15 CL-19). A setting verified off does not clear it: the supplier can change a setting, and only the contract binds it.

The contract carries the term barring training on the organisation's data.

H-3

Personal data would go in, and there are no processor terms (P06 clause CL-06).

Processor terms are signed, and the Data Protection Officer's advice is recorded.

H-4

A provider signal is present: the organisation would build on a supplier's model API, brand, supply or change the purpose of the system (AG-17). The signal is not a classification.

Legal records which role the organisation holds (deployer, provider or other), with its reasoning in A08. If the role is deployer, the tier is set again from the answers. If it is provider or other, the hold clears but the use stays Tier 3: the executive decision-maker decides, with A08 confirmed by Legal.

BC-1

Staff sign in with the work account through single sign-on (P13). No personal or free account. A feature built into a service follows that service's own access rules (P07, P13).

Internal and above

BC-2

The contract bars the supplier from training on the organisation's data (ST-1; P15 CL-19). Where the tool also has a setting that stops training, it is kept off and checked at each review as an added technical control; it never replaces the contract term.

Internal and above

BC-3

The tool and its supplier are in the inventory (A05) and, for Internal and Confidential data, in the supplier register (P06).

Internal and above

BC-4

Confidential data and above only after the supplier's documented decision is Approve or Approve with conditions (P06 TP-05; P15 CL-16).

Confidential and above

BC-5

Personal data only where the tool is approved for it, under processor terms (P06 clause CL-06).

Public

ST-1

The supplier does not use our prompts, files or outputs to train or improve its models or anyone else's, and says so in the contract.

None. Write the term in the contract

ST-2

The supplier processes personal data only on our documented instructions.

P06 clause CL-06

Data protection: processor terms

ST-3

The supplier names the sub-processors it uses, including the provider of the underlying model, and gives notice before it changes them.

P06 clause CL-04

Sub-processors and subcontracting

ST-4

The supplier states where prompts, outputs and logs are stored and processed, and gives notice before it changes the locations.

P06 clause CL-05

Locations of service and data

ST-5

The supplier notifies us of a security incident that exposes our prompts, outputs, files or accounts.

P06 clause CL-07

Security incident notification

ST-6

On exit the supplier returns and deletes our prompts, outputs, logs, uploaded files and any model or index built from our data.

P06 clause CL-13

Data return and deletion

ST-7

The supplier gives notice before it adds or changes an AI feature in a service we already use.

P06 clause CL-20

Notice of material changes

ST-8

The supplier gives us in writing the intended purpose and instructions for use of the AI system (EU AI Act Articles 3(12) and 3(15)).

None. Write the term in the contract

RV-1

Repeat the twelve intake questions with today's answers and run the screen again. A tier higher than the recorded one is a new decision at the higher tier.

RV-2

Check each condition and each dated condition. A condition not met is reported to the approvers, and the tool is switched off if the approval says so.

RV-3

Check the supplier: the decision under P06 still stands, no sub-processor or location has changed without notice (ST-3, ST-4), and no new AI feature has been added (ST-7).

RV-4

Check the people: who still has access, who checks the output, and whether each has completed the AI essentials module.

RV-5

Look at the record of events and complaints about the use since the last review. A mistake that reached a customer, a payment or a person is reported through the incident route (AT-5).

RV-6

Confirm, retire or change. A confirmed use gets a new review date from the tier's months. A retired use is closed in the inventory and its access removed.

EX-1

Find it

Uses are found by the inventory survey, the discovery of systems bought with a card or signed into with a work address (P13 JML-12), a supplier's notice of a new AI feature (ST-7), a manager's question or a staff report. Whoever finds it tells the AI governance lead the same day.

AI governance lead

EX-2

Protect first

Ask the people involved to stop putting in anything above Public. If the use is a Stop (a prohibited practice, or special category data going in), stop it now: do not wait for the working days in the next step. Tell Legal or the Data Protection Officer the same day.

AI governance lead;

Data Protection Officer;

Legal

EX-3

Find out what went in

Ask the users what they put in, over what period, through which account. Compare the answer with the approved list. If Internal or Confidential information went into a tool not approved for it, or personal data into a tool not approved for it, report it as an event (AT-1 to AT-3). Nobody is blamed for a report made in good faith (P09 IM-01).

AI governance lead;

AI use owner

EX-4

Decide whether it can go on

Answer the twelve questions as the use runs today, and mark the use as already running. The screen sets the tier as for a new use. A hold becomes a condition with a date. If the tool is not on the approved list, or the use needs data above what the list allows, stop the use within the working days AG-02 sets, and help the people to a tool they may use.

AI governance lead;

AI use owner

EX-5

Approve with dated conditions, or stop

A use that can go on under the tier's approvers is approved with conditions. Write each dated condition with the date and the person who answers for it, and say in the decision that the tool is switched off if the date is missed. A use that cannot go on is Stopped, with its reason.

AI use owner;

Head of Information Security;

Executive decision-maker

EX-6

Assess within the transition

At Tier 2 or Tier 3 the assessment is due within the calendar days AG-09 sets after the inventory closed. Enter the due date in the inventory (the inventory check IV-08 looks for it) and keep to it.

Head of Information Security

A tool that is not on the approved list, with Internal data going in

Protect first (EX-2). Stop within [[3]] working days (AG-02), report as an event (AT-1), help the people to an approved tool. The use may be proposed afresh.

An approved supplier added an AI feature to software the organisation already uses

Treat the feature as a new use. Screen it. If it can go on, approve it with conditions and a date for the assessment.

A use that matches a prohibited practice or takes special category data

Stop it now (EX-2). Legal is told for a prohibited practice. For special category data report at once and tell the Data Protection Officer the same day (AT-3). There is no dated condition.

IC-A

Draft and summarise internal text with the approved office assistant

Tier 1

AI governance lead decides within [[5]] working days. Review every [[12]] months

None

AG-09

IC-B

Rewrite a published product description in a free public tool

Tier 1

AI governance lead decides within [[5]] working days. Review every [[12]] months

None

AG-09

IC-C

Staff shift rota summarised, with names

Tier 2

AI governance lead and Head of Information Security decide within [[15]] working days. Review every [[6]] months

None

AG-03

IC-D

Customer price lists compared by an AI tool

Tier 2

AI governance lead and Head of Information Security decide within [[15]] working days. Review every [[6]] months

None

AG-03

IC-E

Confidential contract text, but the supplier has not agreed to bar training

Tier 2

AI governance lead and Head of Information Security decide within [[15]] working days. Review every [[6]] months

Hold H-2

AG-03

IC-F

Supplier with no assessment yet, Internal data, a new use

Tier 1

AI governance lead decides within [[5]] working days. Review every [[12]] months

Hold H-1

AG-11

IC-G

The same, but the use is already running

Tier 1

AI governance lead decides within [[5]] working days. Review every [[12]] months

Dated condition H-1

AG-11

IC-H

A forecast that sets reorder quantities for three warehouses

Tier 2

AI governance lead and Head of Information Security decide within [[15]] working days. Review every [[6]] months

None

AG-10

IC-I

A chatbot that answers customers on its own

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

None

AG-16

IC-J

An assistant that drafts replies for agents to check before sending

Tier 2

AI governance lead and Head of Information Security decide within [[15]] working days. Review every [[6]] months

None

AG-03

IC-K

A tool that ranks job applicants

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

None

AG-06

IC-L

Scoring staff performance from call recordings

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

None

AG-06

IC-M

A video-call tool that reports the mood of the team

Stop

Not started. Legal is told

None

AG-05

IC-N

Occupational health notes summarised by an assistant

Stop

Not started. The data protection officer is told; change the use

None

AG-03

IC-O

We would sell our own assistant, built on a supplier's model, to customers

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

Hold H-4

AG-17

IC-P

Staff would use an assistant we build on a supplier's model API, with Internal data (Legal has not recorded the role)

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

Hold H-4

AG-17

IC-Q

The same assistant after Legal records the deployer role

Tier 1

AI governance lead decides within [[5]] working days. Review every [[12]] months

None

AG-09

IC-R

An assistant we build on a supplier's model API for customers, Legal has recorded the role, and the answers give Tier 2

Tier 2

AI governance lead and Head of Information Security decide within [[15]] working days. Review every [[6]] months

None

AG-08

IC-S

A feature built on a supplier's model API and already running before the policy, role not yet recorded

Tier 1

AI governance lead decides within [[5]] working days. Review every [[12]] months

Dated condition H-4

AG-09

IC-T

The same new assistant, but Legal records the provider role: H-4 clears as a hold, and the use stays Tier 3

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

None

AG-17

IC-U

The same new assistant, but Legal records another role: H-4 clears as a hold, and the use stays Tier 3

Tier 3

Executive decision-maker decides within [[30]] working days. Review every [[6]] months

None

AG-17

AI-001

Office suite assistant

Tier 1

27 Aug 2026

1 Sep 2026

3 working days

2 Sep 2026

4 working days from request

Approved with conditions

2 Sep 2027

AI-002

Warehouse system forecasting module

Tier 2

17 Sep 2026

21 Sep 2026

2 working days

28 Sep 2026

7 working days from request

Approved with conditions

28 Mar 2027

AI-003

Customer-service chatbot

Tier 3

9 Sep 2026

11 Sep 2026

2 working days

29 Sep 2026

14 working days from request

Held

Hold due to clear 1 Dec 2026

None set

AI-004

Design tool image generator

Tier 1

16 Sep 2026

17 Sep 2026

1 working day

23 Sep 2026

5 working days from request

Approved with conditions

23 Sep 2027

AI-005

Free public chat assistant

Tier 1

18 Sep 2026

18 Sep 2026

0 working days

21 Sep 2026

1 working day from request

Stopped

None set

AI-006

Security monitoring service

Tier 2

15 Sep 2026

21 Sep 2026

4 working days

28 Sep 2026

9 working days from request

Approved with conditions

28 Mar 2027

AI-007

Recruitment system ranking add-on (proposed)

Tier 3

17 Sep 2026

22 Sep 2026

3 working days

22 Sep 2026

3 working days from request

Declined

None set

AI Acceptable Use & Security Policy

The rules AG-01 to AG-22 this procedure runs, including the prohibitions and the data rules

AI System & Use Case Inventory

Where every use and its tier, approver, conditions, holds and review date are recorded

AI Use Case Risk & Impact Assessment

The assessment for Tier 2 and Tier 3: six factors, rated before and after controls

AI Tool Approval & Data Permission Matrix

The approved list: each tool's state, data levels and conditions, and the terms to ask of a supplier

EU AI Act Role & Risk Classification Worksheet

Legal's worksheet for Tier 3 and for any request that might touch an Annex III area or a prohibited practice

Third-Party Security Risk Management pack (P06)

The supplier assessment (TP-01, TP-05, TP-09) and the contract clauses the supplier terms point to

Security Incident Management pack (P09)

The incident route (IM-01), the screening of declared incidents (IM-06) and the playbooks

ISO/IEC 27001:2022

Annex A 5.8 — Information security in project management

Stages 1 to 4: information security is considered when a use is proposed and approved, before it starts

ISO/IEC 42001:2023

Clause 6.1 — Actions to address risks and opportunities

Stages 2 and 3: actions to address risks and opportunities, by tier

ISO/IEC 42001:2023

Clause 6.1.4 — AI system impact assessment

Stage 3: the impact assessment of a Tier 2 or Tier 3 use

ISO/IEC 42001:2023

Annex A 5.2 — AI system impact assessment process

Stage 3: the process that decides when and how a use is assessed

ISO/IEC 42001:2023

Annex A 9.2 — Processes for responsible use of AI systems

Stages 1 to 7: a defined process for the responsible use of an AI system

NIST CSF 2.0

GV.RM-03 — “Cybersecurity risk management activities and outcomes are included in enterprise risk management processes”

Stages 2 to 4: AI risk handled through the organisation's risk management, with the executive decision-maker deciding the highest-risk uses

EU AI Act — Regulation (EU) 2024/1689

Article 3(4) — “deployer” means “a natural or legal person, public authority, agency or other body using an AI system under its authority except where the AI system is used in the course of a personal non-professional activity”

Scope and Stage 3: the organisation is looking at the deployer test for the AI it uses

EU AI Act — Regulation (EU) 2024/1689

Article 5(1), points (a) to (h) — prohibited, from 2 February 2025: AI that deploys subliminal, purposefully manipulative or deceptive techniques that materially distort behaviour and cause significant harm (a); exploits vulnerabilities due to age, disability or a specific social or economic situation (b); evaluates or classifies people by social behaviour or personal characteristics, with detrimental treatment in unrelated contexts or that is unjustified or disproportionate (c); assesses the risk of a person committing a criminal offence based solely on profiling or personality traits (d); creates or expands facial recognition databases by untargeted scraping of facial images from the internet or CCTV (e); infers emotions in the areas of workplace and education institutions, except for medical or safety reasons (f); categorises people by biometric data to infer race, political opinions, trade union membership, religious or philosophical beliefs, sex life or sexual orientation (g); and, with narrow exceptions, real-time remote biometric identification in publicly accessible spaces for law enforcement (h). Regulation (EU) 2026/1744 adds points (ba) and (bb) (AI that generates or manipulates realistic intimate images of an identifiable person without consent, and AI that generates or manipulates child sexual abuse material), with Article 5(1a) and (1b), from 2 December 2026

Stage 2: the Stop for a prohibited practice, using the checklist PP-01 to PP-10

EU AI Act — Regulation (EU) 2024/1689

Article 6(2) — in addition to systems that are safety components of products under Annex I, “AI systems referred to in Annex III shall be considered to be high-risk”

Stage 2: an Annex III area sends a request to Tier 3 and to Legal

EU AI Act — Regulation (EU) 2024/1689

Article 6(3) — an Annex III system is not high-risk where it “does not pose a significant risk of harm to the health, safety or fundamental rights of natural persons, including by not materially influencing the outcome of decision making”, which applies where it is intended to perform a narrow procedural task, improve the result of a previously completed human activity, detect decision-making patterns without replacing or influencing the human assessment, or perform a preparatory task; “an AI system referred to in Annex III shall always be considered to be high-risk where the AI system performs profiling of natural persons”. The provider documents the assessment before the system is placed on the market (Article 6(4))

Stage 3: Legal's reading of a request that touches an Annex III area, including profiling

EU AI Act — Regulation (EU) 2024/1689

Article 25(1) — a distributor, importer, deployer or other third party “shall be considered to be a provider of a high-risk AI system” and subject to the provider’s obligations where it puts its name or trademark on a high-risk system already on the market, makes a substantial modification to a high-risk system that remains high-risk, or modifies the intended purpose of a system that was not high-risk so that it becomes high-risk under Article 6

Stage 2: a provider signal (building on a model API, branding, change of purpose) sends the request to Legal first

EU AI Act — Regulation (EU) 2024/1689

Article 26(1) — deployers of high-risk AI systems “shall take appropriate technical and organisational measures to ensure they use such systems in accordance with the instructions for use accompanying the systems”. The obligations of Article 26 applies from 2 December 2027 to systems classified as high-risk under Article 6(2) and Annex III, and from 2 August 2028 to those under Article 6(1) and Annex I (Article 113, third paragraph, point (c), as amended by Regulation (EU) 2026/1744)

Stage 3: the deployer duties for a high-risk system that the request might bring

EU AI Act — Regulation (EU) 2024/1689

Article 26(7) — before putting into service or using a high-risk system at the workplace, deployers who are employers “shall inform workers’ representatives and the affected workers that they will be subject to the use of the high-risk AI system”

Stage 3: informing workers and their representatives, through HR

EU AI Act — Regulation (EU) 2024/1689

Article 26(9) — where applicable, deployers of high-risk systems use the information the provider gives under Article 13 to carry out their data protection impact assessment under Article 35 of Regulation (EU) 2016/679

Stage 3: the data protection impact assessment, using the provider's information

EU AI Act — Regulation (EU) 2024/1689

Article 50(1) — providers ensure that AI systems intended to interact directly with natural persons are designed so that the persons are informed that they are interacting with an AI system, “unless this is obvious from the point of view of a natural person who is reasonably well-informed, observant and circumspect”

Stage 6: telling people that they are dealing with an AI system

EU AI Act — Regulation (EU) 2024/1689

Article 113, as amended — the Regulation applies from 2 August 2026, except that Chapters I and II (including Articles 4 and 5) apply from 2 February 2025, with Article 5(1) points (ba) and (bb) and Article 5(1a) and (1b) applying from 2 December 2026; Chapter III Section 4, Chapter V (general-purpose AI models), Chapter VII and Chapter XII and Article 78 apply from 2 August 2025, except Article 101; and Chapter III Sections 1, 2 and 3, except Article 6(5), apply from 2 December 2027 as regards systems classified as high-risk under Article 6(2) and Annex III, and from 2 August 2028 as regards those under Article 6(1) and Annex I. Article 50 is not listed in an exception and so applies from 2 August 2026

The legal note and the dates: when each part of the Act applies after Regulation (EU) 2026/1744

EU AI Act — Regulation (EU) 2026/1744 (Digital Omnibus on AI)

Regulation (EU) 2026/1744 of 8 July 2026 amending Regulations (EU) 2024/1689, (EU) 2018/1139 and (EU) 2023/1230 as regards the simplification of the implementation of harmonised rules on artificial intelligence (Official Journal, 24 July 2026); in force on the third day after publication (its Article 4), 27 July 2026. Its Article 1 amends Regulation (EU) 2024/1689

The legal note: the amending Regulation, in force from 27 July 2026

EU AI Act — Commission guidelines (non-binding)

Draft Commission guidelines on the classification of high-risk AI systems, published 19 May 2026 for targeted consultation (general principles; Annex I; Annex III); a draft on 8 October 2026, not adopted

Stage 3: the draft guidelines on classification. They are a draft, so Legal confirms each reading

NIST AI RMF 1.0

GOVERN 1.3 — “Processes, procedures, and practices are in place to determine the needed level of risk management activities based on the organization’s risk tolerance.”

Stages 2 and 4: the level of risk management follows the organisation's risk tolerance

NIST AI RMF 1.0

MAP 3.5 — “Processes for human oversight are defined, assessed, and documented in accordance with organizational policies from the GOVERN function.”

Stage 4 and conditions: human oversight defined for each use

NIST AI RMF 1.0

MANAGE 1.1 — “A determination is made as to whether the AI system achieves its intended purposes and stated objectives and whether its development or deployment should proceed.”

Stage 4: a decision whether the use should proceed

GDPR — Regulation (EU) 2016/679

Article 22(1) — “The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her.” Article 22(3) requires, for the cases in Article 22(2), points (a) and (c), at least the right to obtain human intervention, to express a point of view and to contest the decision

Stage 2: a use that decides or shapes a decision about a person is Tier 3

GDPR — Regulation (EU) 2016/679

Article 28(3) — a binding contract with the processor, covering documented instructions, confidentiality, security, sub-processors, assistance, return or deletion of data, and audits

Hold H-3 and ST-2: processor terms

GDPR — Regulation (EU) 2016/679

Article 35(1) — where a type of processing, in particular using new technologies, is likely to result in a high risk to the rights and freedoms of natural persons, the controller carries out “prior to the processing” an assessment of the impact on the protection of personal data; a single assessment may address a set of similar processing operations with similar high risks

Stage 3: the data protection impact assessment

GDPR — Regulation (EU) 2016/679

Article 35(3) — a data protection impact assessment is required in particular for a systematic and extensive evaluation of personal aspects of natural persons based on automated processing, including profiling, on which decisions with legal or similarly significant effects are based; for processing on a large scale of special categories of data; and for systematic monitoring of a publicly accessible area on a large scale

Stage 3: the cases where an assessment is required

NIS2 — Directive (EU) 2022/2555

Article 21(2)(d) — “supply chain security, including security-related aspects concerning the relationships between each entity and its direct suppliers or service providers”

Stage 3 and supplier terms: supply chain security

AI use

One way the organisation uses an AI system, for one purpose, with one owner: summarising meeting notes with the office assistant, or forecasting stock with the warehouse system. One tool can have several uses. The inventory lists uses.

Approved tool

An AI tool on the approved list (A07) with a state of Approved or Approved with conditions, and the data levels it may take.

Tier

How much checking a proposed use needs before it starts: Tier 1 a screen, Tier 2 an assessment, Tier 3 an assessment and an executive decision. A fourth outcome, Stop, means the use is not allowed.

Unapproved use

Any use of an AI tool for work that is not on the approved list, or with data above what the list allows. Sometimes called shadow AI.