// Trust & Security Report
Allen Institute for AI (Ai2)
Semantic Scholar is a free, AI-powered academic literature search and discovery tool operated as a Research and Product Development team within Ai2 (nonprofit). Sub-products include the public web app, Semantic Reader (beta augmented reader), and the Semantic Scholar Academic Graph API for developers.
Certifications held
1
Maturity
Startup
Trains on your data
Unknown
Trust center
No
// Certification ledger
Each held certification is backed by a verbatim quote from the vendor's own trust or security page. “Not confirmed” means we could not verify it publicly, not that the vendor lacks it.
Verify on allenai.org“Provided that you reside in the European Economic Area or the United Kingdom, or that GDPR otherwise applies, you have the following data subject rights in accordance with the EU GDPR...”
> Show 8 unconfirmed / not-held certifications
source: allenai.orgno public evidence: no trust center, security page, or SOC 2 report/badge found on semanticscholar.org or allenai.org after checking the homepage, about page, privacy policy, and API license
source: allenai.orgno public evidence: not mentioned in Ai2's privacy policy, Semantic Scholar about page, or API license agreement
source: allenai.orgno public evidence found on vendor domain; Ai2 emphasizes open-source transparency practices but no formal ISO 42001 certification claim located
source: allenai.orgno public evidence; not mentioned anywhere on semanticscholar.org or allenai.org, and the product does not handle protected health information
source: semanticscholar.orgnot applicable / no public evidence; Semantic Scholar is a free service with no payment processing on the platform
// Privacy & AI training
Trains on customer data
Not stated
Data processing agreement
Not offered
Data region
United States (privacy policy states: 'Your information will be stored and processed by Ai2 in the United States,' with Standard Contractual Clauses / Data Privacy Framework safeguards for EEA transfers)
For general use of Semantic Scholar (search, browsing, author pages) there is no indication of training on private user data; the underlying corpus is public/licensed academic literature. For interactive AI demo features, the privacy policy states: 'we process any personal data included in those prompts exclusively based on your consent and to improve our AI demonstrations...' indicating consent-based use of prompt data to improve demo features specifically, not a blanket statement about all products or an explicit opt-out mechanism being described.
// Security controls
Encryption in transit/at rest
Not specified. Privacy policy only states a general commitment: 'We work hard to protect information you provide from loss, misuse, and unauthorized access or disclosure,' with no detail on encryption standards.
allenai.orgVulnerability disclosure
A security.txt file is published at semanticscholar.org/.well-known/security.txt listing security@allenai.org as the security contact (expires 2027-03-01), indicating a basic responsible-disclosure channel exists.
semanticscholar.orgData retention
Personal information retained only as long as necessary for the purposes for which it was collected, per the privacy policy.
allenai.orgThird-party/legal contact for security safeguards
Further information on safeguards can be requested by email at legal@allenai.org.
allenai.org// Products & data scope
data: Public academic paper metadata, user account info (email, name) if registered, optional claimed author-page data
No paid tier; free for all users, no enterprise SLA or trust center.
data: Public/licensed academic literature data (S2 Data under CC BY-NC / ODC-BY style licenses); API key usage logged
Governed by a separate API License Agreement; rate-limited, attribution required, liability capped at $100.
data: Same public paper corpus, contextual AI annotations for select papers
Beta feature, limited to select papers; no distinct security disclosures found.
// What to watch
- No trust center, no SOC 2/ISO 27001/HIPAA/PCI evidence found anywhere on the vendor's own domains (semanticscholar.org, allenai.org) despite searching the privacy policy, about page, API license, and general web search — consistent with a free nonprofit research tool, not an over-claim, but should be listed with an explicit 'no certifications published' caveat rather than left blank.
- AI-training data usage is ambiguous outside the specific 'AI demonstrations' consent clause; no explicit statement covers whether the core search/API corpus processing trains any Ai2 models on user-submitted (vs. public paper) data, and no opt-out mechanism is described.
- Identity note: Semantic Scholar has no separate legal entity or privacy policy of its own; all privacy/legal terms are Ai2's, so certifications or incidents attributed to 'Ai2' broadly should be understood as covering Semantic Scholar too, but any AWS/cloud-host certifications (not found in this search) would belong to the host, not Ai2/Semantic Scholar, and must not be conflated if encountered later.
// At a glance
Pricing model
Free (nonprofit-funded, no paid tiers identified)
Self-hostable
No
// How we verified this
Every certification marked HELD is confirmed against a verbatim quote on Allen Institute for AI (Ai2)'s own trust, security, or privacy pages. We reject certifications claimed only on third-party aggregators, on a cloud host's behalf, or by a similarly named company.
Last verified 2026-07-09. Compliance changes over time. Always confirm directly with the vendor before relying on any certification for a purchasing or compliance decision.
allenai.org