Joint research study
SecureSpace and a collaborator define a question, method, responsibilities, outputs, and review process.
Research Collaboration
Bring the question worth studying.
SecureSpace welcomes structured proposals from universities, companies, researchers, students, and security practitioners.
We focus on work that connects rigorous inquiry with real intelligent systems.
Proposal quality
A strong proposal begins with a specific question.
The strongest research proposals are not simply requests to collaborate on AI. They define a meaningful security problem, explain why it matters, describe what is already known, identify the proposed method, and clarify what each collaborator may contribute.
SecureSpace is particularly interested in questions whose answers could improve system design, security practice, evaluation, governance, evidence, or future infrastructure.
Research question
System or context
Why the problem matters now
Existing work
Proposed methodology
Required data or infrastructure
Expected contribution from each party
Intended outcome
Publication preference
Confidentiality expectations
Timeline
Funding requirements, if any
Collaboration areas
SecureSpace is most interested in questions that connect security research to system reality.
Agent security
Prompt injection and instruction integrity
Agent identity and access
Tool-use security
Retrieval and provenance
Memory and context security
AI application security
Secure AI development
API security
Cloud and AI infrastructure
Runtime monitoring
Human oversight
AI governance
Security evaluation
Benchmark design
Evidence and auditability
Privacy
Security architecture
Digital trust
Formats
Collaboration can take several shapes, depending on the question.
Applied research pilot
Research is conducted around a real system, product, workflow, or enterprise constraint.
University collaboration
Faculty, laboratories, researchers, and students work with SecureSpace on defined security questions.
Enterprise research programme
An organisation collaborates on a difficult security question grounded in operational systems and constraints.
Evaluation or benchmark development
Collaborators develop a method for testing, comparing, or evaluating an AI-security problem.
Technical workshop
A collaborative workshop is used to clarify a research question, threat model, or evaluation approach.
Joint publication
Publication may be considered where the evidence, permissions, safety requirements, and authorship expectations are clear.
Private research engagement
Some work may remain confidential and produce internal findings rather than public material.
SecureSpace contribution
Support depends on fit, safety, capacity, and formal agreement.
Availability of support, infrastructure, funding, data, or product access is never guaranteed.
Security research framing
AI-security context
Threat modelling
Applied system knowledge
Security architecture
Evaluation methodology
Engineering collaboration
Product and infrastructure context
Enterprise problem framing
Research review
Publication support
Access to selected non-sensitive research artefacts
Potential connection to Mintos AI product questions
Review process
How research proposals move from idea to possible collaboration.
01
Submission
The applicant provides the research question, background, proposed collaboration, expected contribution, timeline, and relevant material.
02
Initial relevance review
SecureSpace reviews whether the proposal aligns with current research themes and capacity.
03
Feasibility review
The proposal is assessed for scope, access, methodology, data requirements, safety, confidentiality, resources, and expected value.
04
Introductory discussion
Selected proposals may proceed to a conversation to clarify the problem and collaboration structure.
05
Internal decision
SecureSpace determines whether to decline, request more information, explore further, or propose a formal structure.
06
Collaboration agreement
Any accepted collaboration requires separate agreement covering responsibilities, confidentiality, intellectual property, publication, data handling, and funding where relevant.
Confidentiality
Do not send sensitive material through the initial application.
Applicants should not submit trade secrets, credentials, private keys, regulated personal information, classified material, confidential customer data, or sensitive system details through the initial form unless SecureSpace has explicitly agreed to receive them.
A mutual NDA or separate secure exchange may be considered later where appropriate.
What submission does not guarantee
A response beyond acknowledgement
A meeting
A partnership
Funding
Product access
Infrastructure access
Dataset access
Internship
Employment
Publication
Authorship
Detailed feedback
Mintos AI access
Questions teams usually ask
Who can apply?
Universities, companies, independent researchers, students, security practitioners, public-interest organisations, and relevant technical teams can submit proposals.
Do applicants need to represent a university?
No. University proposals are welcome, but independent researchers and companies can also apply.
Can students apply?
Yes, especially where a project is supervised or clearly scoped. Student submission does not guarantee a programme, internship, or funding.
Does SecureSpace fund research?
Funding is not guaranteed. Any funding arrangement would require separate review and written agreement.
Can proposals remain confidential?
Potentially. Confidentiality terms must be agreed separately before sensitive information is shared.
Can the work be published?
Publication may be possible where evidence, safety, permissions, authorship, confidentiality, and disclosure requirements are clear.
Does collaboration include Mintos AI access?
No access should be assumed. Any product, infrastructure, or data access must be explicitly agreed.
Continue through the Research section
Next step
Bring a question worth investigating.
Prepare the research problem, current evidence, proposed method, collaboration request, and the contribution you or your organisation can make.