Annex XIII AI Act
class="oj-doc-ti" id="d1e38-144-1">IARSCRÍBHINN XIII
Official text
id="anx_XIII"> IARSCRÍBHINN XIII Critéir le haghaidh samhlacha intleachta saorga ilchuspóireacha a ainmniú a bhfuil riosca sistéamach acu dá dtagraítear in Airteagal 51 Chun críche a chinneadh go bhfuil cumais nó tionchar atá coibhéiseach leo sin a leagtar amach in Airteagal 51(1), pointe (a), ag samhail intleachta saorga ilchuspóireach, cuirfidh an Coimisiún na critéir seo a leanas san áireamh:
(a) líon pharaiméadair na samhla;
(b) cáilíocht nó méid an tacair sonraí, mar shampla tomhaiste trí chomharthaí;
(c) an méid ríomha a úsáidtear le haghaidh oiliúint a chur ar an tsamhail, arna thomhas in oibríochtaí snámhphointe nó arna léiriú le meascán d’athróga eile mar shampla costas measta na hoiliúna, an t-am a mheastar a bheidh ag teastáil le haghaidh na hoiliúna, nó ídiú measta fuinnimh le haghaidh na hoiliúna;
(d) módúlachtaí ionchuir agus aschuir na samhla, mar shampla téacs go téacs (samhlacha móra teanga), téacs go híomhá, ilmhódúlacht, agus tairseacha úrscothacha chun cumais ardtionchair a chinneadh le haghaidh gach modhúlachta, agus cineál sonrach na n-ionchur agus na n-aschur (e.g. seichimh bhitheolaíocha);
(e) na tagarmharcanna agus meastóireachtaí ar chumais na samhla, lena n-áirítear líon na gcúraimí gan oiliúint bhreise, inoiriúnaitheacht chun cúraimí nua ar leith a fhoghlaim, leibhéal a neamhspleáchais agus a hinscálaitheachta, na huirlisí a bhfuil rochtain aici orthu;
(f) an bhfuil tionchar mór aici ar an margadh inmheánach mar gheall ar a rochtain, rud a thoimhdeofar nuair a bheidh sí curtha ar fáil do 10 000 úsáideoir gnó cláraithe ar a laghad atá bunaithe san Aontas;
(g) líon na n-úsáideoirí deiridh cláraithe.
ELI: http://data.europa.eu/eli/reg/2024/1689/oj ISSN 1977-0839 (electronic edition)
////////////////////////$(document).ready(function(){generateTOC(true,'', 'Barr','false');scrollToCurrentUrlAnchor();});
Source: EUR-Lex, Regulation (EU) 2024/1689 — text reproduced verbatim.
📬 AI Act Weekly
Get the most important AI Act developments in your inbox every week.
SubscribeFrequently asked questions
What does Annex XIII of the AI Act regulate?
Annex XIII describes the criteria the European Commission uses to determine whether a GPAI model poses systemic risk, under Article 51. These criteria determine whether additional obligations apply.
Which criteria determine if a GPAI model has systemic risk?
Seven criteria: (a) number of parameters, (b) dataset quality/size, (c) amount of compute for training, (d) input/output modalities and state-of-the-art thresholds, (e) benchmarks and evaluations, (f) internal market impact (10,000+ business users), and (g) number of registered end-users.
Is there a threshold for the number of parameters?
Annex XIII lists number of parameters as a criterion (point a), but does not set a fixed threshold. Article 51(1)(a) does mention a threshold of 10^25 FLOP of compute for training as a presumption of systemic risk.
When is a model presumed to have systemic risk?
Article 51(2) provides that a model is presumed to have systemic risk if cumulative compute for training exceeds 10^25 FLOP. The Commission may adjust this threshold through delegated acts.
Does the number of users count for systemic risk?
Yes, point f states that high impact on the internal market is presumed when the model has been made available to at least 10,000 registered business users in the EU. Point g also references the number of end-users.
Are multimodal models more likely to be systemic risk?
Annex XIII, point d, specifically mentions input/output modalities (text-to-text, text-to-image, multimodal) as a criterion, with state-of-the-art thresholds per modality. Multimodal capabilities can therefore contribute to the classification.
Who determines whether a model has systemic risk?
The European Commission determines this based on the criteria in Annex XIII and Article 51. The provider may also self-declare that their model has systemic risk. The 10^25 FLOP threshold creates a legal presumption.
What are the consequences of systemic risk classification?
Providers of GPAI models with systemic risk must comply with additional obligations: model evaluation, adversarial testing, incident reporting to the AI Office, and adequate cybersecurity protection (Article 55).