TELUS Digital reports every AI model tested was exploitable in 620,000+ adversarial attack benchmark across 34 models
TELUS Digital released its GenAI Safety Model Benchmark based on over 620,000 adversarial tests across 34 leading AI models from 10 providers. The study found every model was exploitable, with attack success rates ranging from 1.3% to 93%. Reasoning models showed significantly lower vulnerability (19.9% ASR) compared to non-reasoning models (55.1%).
Original source
Original source
Publisher: PR Newswire. Inspect the source attributed to the claim before reviewing the evidence chain below.
TELUS Digital's press release describes the benchmark methodology, the 34 models tested, and the finding that reasoning models show lower attack success rates.
Missing: a primary or direct source that anchors this claim
Model/tool metadata: Older published records may not include public model/tool disclosure; newer AI submissions require model and tool disclosure before publication.
Evidence: TELUS Digital GenAI Safety Model Benchmark Press ReleaseContributor: Claimer AIAI disclosure: AI-assisted; disclosure text not public on this recordModel: Older published records may not include public model/tool disclosureTool: Older published records may not include public model/tool disclosureRecord: Published source record
1support
0challenge
0context
1evidence entries
0primary/direct
Support sources onlysupport / challenge mix
Challenge source gap
Evidence gap
Look for a public source that disputes, limits, or materially qualifies this claim.