Skip to content
TechCrunch·

🔒AI Labs Fear Being Seen as Trojan Horses

Your AI Model Might Be Spying on You

TL;DR

AI labs selling proprietary models are concerned about being seen as Trojan horses, potentially giving away sensitive business information to competitors. Microsoft warns users pay twice: with money and data.

AI labs selling proprietary models are worried they're becoming Trojan horses for competitors, potentially leaking valuable business insights. This is a big deal if you rely on AI for competitive advantage or handle sensitive data. Enterprises teaching these models about their businesses risk giving away trade secrets to the model makers. Open-source models installed on-premises can now do almost 90% of what proprietary models do at lower costs, making them an attractive alternative.

AI Labs Fear Being Seen as Trojan Horses — TechCrunch

Key Points

1

Startups and enterprises using AI from OpenAI or Anthropic risk giving away valuable trade secrets (3).

2

Microsoft CEO warns that AI users pay twice: once in dollars and again with valuable business data (4).

3

Enterprises teaching proprietary models about their businesses can be used by competitors to gain an edge (5).

4

Open-source models are now doing almost 90% of what proprietary models do, at a lower cost (12).

5

On-premise open-source models are becoming increasingly popular among enterprises as they control their data and reduce risk (13)

Why It Matters

If you're using AI to gain competitive advantage or handle sensitive business information, proprietary models might be leaking your secrets. Open-source models installed on-premises can do almost 90% of what proprietary models do at a lower cost, making them an attractive alternative for those concerned about data security and control.

data-securityopen-sourceproprietary-modelsenterprise-ai

Frequently Asked Questions

Why does this matter?

If you're using AI to gain competitive advantage or handle sensitive business information, proprietary models might be leaking your secrets. Open-source models installed on-premises can do almost 90% of what proprietary models do at a lower cost, making them an attractive alternative for those concerned about data security and control.

What happened?

AI labs selling proprietary models are concerned about being seen as Trojan horses, potentially giving away sensitive business information to competitors. Microsoft warns users pay twice: with money and data.

Comments

Subscribe to join the conversation...

Be the first to comment

Enjoyed this article?

Get it daily. 7am. Free. Reads in 5 minutes.