# Defense AI Faces Data Manipulation Risks

An article on defense AI security identifies data poisoning and indirect prompt injection as risks to AI systems and highlights data provenance verification, data quality checks, and resilient-by-design systems as relevant safeguards. It reports no solicitation or contract award, but the risks and controls offer contractors working in defense AI, data infrastructure, and cybersecurity concrete themes for solution development and customer discussions. The 2027 Defense R&D Summit is scheduled for February 4, 2027, providing a forum to discuss related mission needs.

- Contractors can use provenance verification, data quality checks, and system resilience as specific capabilities to address in defense AI technical discussions and solution planning; the signal does not establish these as mandatory requirements.
- The article identifies no funded opportunity or procurement timeline, so it should be treated as market intelligence rather than an active bid.
- The 2027 Defense R&D Summit takes place February 4, 2027, 99 days from October 28, 2026, and offers an opportunity to discuss defense AI security needs.

**Jurisdictions:** federal
**Industries:** Defense & Military, Information Technology
**Topics:** Cybersecurity, Artificial Intelligence
**Published:** October 27, 2026

### Government Entities
- Department of War (DoW)
- Cybersecurity and Infrastructure Security Agency (CISA)
- Intelligence Advanced Research Projects Activity (IARPA)

### Vendors
- Palo Alto Networks ()
- Anthropic ()
- OpenAI ()
- Google ()

### Sources
- [Securing Defense AI Against Data Manipulation](https://www.executivegov.com/articles/ai-dow-defense-data-poisoning-threat) - ExecutiveGov