Karsten Wenzlaff, Advisor
August 26th, 2025
Karsten Wenzlaff is a Crowdfunding and Fintech Expert from Germany. He is Secretary-General of Digit[...]
Prompt Injection Attacks in Slack’s AI Integration
AI Risks | Aug 22, 2024
Image: Freepik/rawpixel.com
Bad actors can access private channel slack data using prompt injection attacks
Security experts, PromptArmor, found a serious flaw in Slack's AI integration in August 2024 that may allow private channel communications to be exposed using a sophisticated method called prompt injection. By including a certain set of written instructions in public channels, bad actors can control Slack's AI and potentially cause private channel data breaches without having access. Companies and individuals should be careful of the private data they share on these platforms.
Prompt Injection Exploit
PromptArmor discovered that Slack's AI can access both public and private channels, and be tricked into sharing private conversations containing sensitive information. The attacker doesn't even need to be present in the channel that is being targeted. All they have to do is insert malicious instructions into public channels that the AI can access, and the AI will be tricked into sharing private information like API keys.
See: Meta Agrees to Pay $14B to Texas for Data Privacy Settlement
By extending AI's search capabilities to files and direct messages on August 14, 2024, Slack unintentionally raised the risk of quick injection attacks. These days, malicious code that starts the exploit can also be found in documents like PDFs.
PromptArmor believes:
"If a user downloads a PDF that has one of these malicious instructions (e.g. hidden in white text) and subsequently uploads it to Slack, the same downstream effects of the attack chain can be achieved."
Corporate Data at Risk
In 2020, Slack had over 12 million active daily users, where most of them use the platform for essential business communication. Sensitive information, such as financial information, proprietary data, and private chats, could be exposed due to this vulnerability. With such extensive adoption, this data breach risk is highly concerning.
Slack's Response
Slack quickly released a patch on August 20, 2024 in attempt to fix the attack saying that enabling AI's search capacity across open channels was deliberate (trying to downplay the seriousness of the vulnerability. However, even with the patch in place, questions remain on how well Slack's AI governance can protect user data.
See: Feds Warn Companies to Not Delete Slack Chats
Akhil Mittal, Senior security consulting manager at Synopsys said:
"This vulnerability reveals gaps in AI’s security framework that need urgent addressing, especially as more businesses adopt these tools to manage critical data".
Outlook
Although Slack has fixed the immediate issue, companies still need to be on the lookout and advocate for strong AI security measures. To stop future attacks, regular audits, AI access controls, and employee data security training are essential. 'User beware'
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
 |  |  |
|---|---|---|
 |  |  |
Leave a Reply