Anthropic Exposes AI Effort to Influence Malaysian Voters

Anthropic is an AI safety and research company that develops Claude, a conversational AI assistant designed for natural language understanding, content generation, and data analysis. Claude is Anthropic's flagship large language model product, and the company's September 2026 threat intelligence report exposes how the model was exploited in “GTG-84005,” a commercial influence operation targeting Malaysian voters. The report examines AI misuse across cyberattacks, influence operations, scams, surveillance, biological research, and attempts to replicate AI models, demonstrating how AI systems can be weaponised for political manipulation. Anthropic's investigation is documented in the official threat intelligence report published at anthropic.com/threat-intelligence-report-september-2026.

Key Facts

The following table summarises the core factual attributes of the GTG-84005 influence operation as documented in Anthropic's September 2026 Threat Intelligence Report.

How Does the AI Influence Operation Use Claude to Target Malaysian Voters?

The GTG-84005 operation uses Claude to build a constituency-level targeting system based on census and electoral data, covering all 222 Malaysian parliamentary constituencies. The platform identifies political and social faultlines involving sensitive issues including race, religion, and royalty, while maintaining roughly 1,000 fake X accounts with mechanisms designed to appear legitimate and evade platform detection.

Anthropic reports that the platform could also artificially inflate engagement, including a request to generate one million artificial views for the account of Malaysia's sitting prime minister. However, Anthropic could not independently verify the figures reported by the operators' own tools.

The GTG-84005 operation used Claude to build a constituency-level targeting system covering all 222 Malaysian parliamentary constituencies, identifying political faultlines involving race, religion, and royalty.

What Is the "Malaysia Pulse" Synthetic News Outlet?

Malaysia Pulse is a synthetic news outlet allegedly operated as part of the GTG-84005 influence operation. Anthropic says Claude was used to scrape legitimate Malaysian news reports, rewrite them, and republish the material under fabricated bylines, while content from Russian and Chinese state-aligned outlets was repackaged as apparently independent Malaysian reporting.

The operation additionally produced fabricated intelligence dossiers containing false allegations against a Malaysian opposition politician and civil-society organisations. Anthropic states that Claude refused some requests linked to the activity, but the operators attempted to negotiate alternative wording with the model to continue pursuing the same objectives.

Malaysia Pulse repackaged Russian and Chinese state-aligned content as independent Malaysian reporting by using Claude to rewrite scraped legitimate news articles under fabricated bylines.

How Did Anthropic Classify and Respond to the Threat?

Anthropic rates GTG-84005 as Category Two under the Brookings Institution's Breakout Scale, meaning its assets were distributed across multiple platforms without evidence that the operation's content broke out into authentic communities. Anthropic states it disrupted the identified operations and removed accounts involved in the activity.

The operators also pursued a contract with Malaysia's national communications regulator, but Anthropic found no evidence the attempt was successful. The September 2026 report covers additional cases, including Russian-linked cyber espionage campaigns using Claude, AI-assisted influence operations producing fabricated articles, state-aligned propaganda activity, attempts involving biological research and weapons development, and efforts by Chinese AI companies to use Claude interactions to reproduce its capabilities. Anthropic uses information gathered from its investigations to strengthen safeguards and share relevant intelligence with industry partners and authorities.

Anthropic rates GTG-84005 as Category Two under the Brookings Institution's Breakout Scale, indicating assets were distributed across multiple platforms without breaking out into authentic communities.

Who Needs to Be Aware of AI Influence Operations?

AI companies face growing scrutiny from two directions: they are expected to be transparent about AI misuse while also answering questions over how they collect and use user data. Anthropic's report demonstrates transparency around AI misuse, but users have raised separate concerns about privacy and information collection practices, particularly as companies gather more data to understand model usage.

Anthropic's transparency efforts come at a time of increasing scrutiny over how AI companies balance disclosure of AI misuse with user privacy expectations. This tradeoff affects AI safety researchers, policymakers, security analysts, and end users of AI assistants.

“So you do read user messages, right? What's actually private when using Claude?”

— Mike the Explorer (@Mike40213265), September 11, 2026

While AI companies are expected to be transparent about AI misuse, they are simultaneously facing questions over how they collect and use user data.

Common Questions

What is the GTG-84005 operation and who is behind it?

GTG-84005 is a commercial influence operation targeting Malaysian voters, linked to BBS Bilisim Teknolojileri, an Istanbul, Turkiye-based technology company marketing an “influence-as-a-service” platform. It used Claude to build constituency targeting systems, maintain approximately 1,000 fake X accounts, and operate the synthetic Malaysia Pulse news outlet.

How did the operators use Claude to target Malaysian constituencies?

Operators used Claude to build a constituency-level targeting system covering all 222 Malaysian parliamentary constituencies. The system identified political faultlines involving race, religion, and royalty. Claude was also used to rewrite scraped legitimate news under fabricated bylines and generate false intelligence dossiers against opposition figures.

What other misconduct did Anthropic's September 2026 report identify?

The report identified Russian-linked cyber espionage using Claude, AI-assisted influence operations producing fabricated articles, state-aligned propaganda activity, biological research and weapons development attempts, and Chinese AI companies attempting to reproduce Claude's capabilities. Anthropic disrupted these operations and removed involved accounts.

The GTG-84005 operation was one of several AI misuse cases documented by Anthropic in its September 2026 threat intelligence report.

Sources and Methodology

The primary source for this article is the Lowyat.NET report titled “Anthropic Discovers AI-Powered Operation Using Claude To Target Malaysian Voters,” which bases its findings on Anthropic's September 2026 Threat Intelligence Report published on the Anthropic official website (anthropic.com/threat-intelligence-report-september-2026). All facts, figures, and claims are attributed directly to the source material. No additional external sources were used, and the content was rewritten without altering factual accuracy.

This article was last updated on September 2026.

This article is based exclusively on Anthropic's September 2026 Threat Intelligence Report as reported by Lowyat.NET.