Annex XIII AI Act
class="oj-doc-ti" id="d1e38-144-1">PRILOGA XIII
Official text
id="anx_XIII"> PRILOGA XIII Merila za označitev modelov UI za splošne namene s sistemskim tveganjem iz člena 51 Pri ugotavljanju, ali ima model UI za splošne namene zmogljivosti ali vpliv, enakovredne tistim iz člena 51(1), točka (a), Komisija upošteva naslednja merila:
(a) število parametrov modela;
(b) kakovost ali velikost nabora podatkov, merjeno na primer z žetoni;
(c) količino izračuna, uporabljeno za učenje modela, merjeno z operacijami s plavajočo vejico ali izraženo s kombinacijo drugih spremenljivk, kot so ocenjeni stroški učenja, predvideni čas, potreben za učenje, ali ocenjena poraba energije za učenje;
(d) vhodne in izhodne načine modela, kot so besedilo v besedilo (veliki jezikovni modeli), besedilo v podobo, večmodalnost in najsodobnejši pragovi za določanje zmogljivosti z visoko učinkovitostjo za vsak način, ter posebna vrsta vhodnih in izhodnih podatkov (npr. biološka zaporedja);
(e) merila uspešnosti in ocenjevanja zmogljivosti modela, vključno z upoštevanjem števila nalog brez dodatnega učenja, prilagodljivostjo za učenje novih, različnih nalog, njegovo stopnjo avtonomije in nadgradljivosti ter orodji, do katerih ima dostop;
(f) ali ima velik vpliv na notranji trg zaradi svojega dosega, na katerega se domneva, ko je bil dan na voljo vsaj 10 000 registriranim poslovnim uporabnikom s sedežem v Uniji;
(g) število registriranih končnih uporabnikov.
ELI: http://data.europa.eu/eli/reg/2024/1689/oj ISSN 1977-0804 (electronic edition)
////////////////////////$(document).ready(function(){generateTOC(true,'', 'Na vrh','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).