Microsoft blocks prompts that may lead Copilot AI to generate sexual, violent images

[Trendingger]

 

Microsoft has taken a significant step towards ensuring the responsible use of its artificial intelligence (AI) tool, Copilot. This move comes in response to concerns raised by a staff AI engineer about Copilot’s image generation.

 

The tech giant has started blocking certain terms that were causing its AI tool to generate violent and sexual images. This action was taken after it was discovered that Copilot was generating inappropriate imagery based on these prompts. The terms that are now blocked include “pro choice,” “pro choce” [sic], “four twenty,” and "pro life".

 

Microsoft’s decision to block these terms is an indication of its commitment to adhering to its responsible AI principles. The company has also implemented a warning system for multiple policy violations, which could lead to suspension from the tool. This is a marked change from earlier practices and shows Microsoft’s proactive approach in addressing potential misuse of its AI tool.

 

This incident has sparked a broader conversation about the ethical use of AI and the need for robust guardrails to prevent the generation of harmful content. As AI continues to evolve and become more sophisticated, the challenges associated with its use are also likely to increase. Therefore, it is crucial for tech companies to continuously monitor and make adjustments to their AI tools to mitigate any potential misuse.

 

What are your thoughts on this development? Do you think Microsoft’s actions are sufficient to address the concerns raised about Copilot’s image generation? We’d love to hear your views on this topic.

 

 

Read more:

• https://www.pcmag.com/news/microsoft-starts-blocking-some-terms-that-caused-its-ai-to-create-inappropriate
• https://www.engadget.com/microsofts-copilot-now-blocks-some-prompts-that-generated-violent-and-sexual-images-213859041.html
• https://www.wsj.com/tech/ai/microsoft-engineer-shane-jones-warns-copilot-ai-tool-generates-harmful-images-needs-better-safeguards-659ac2a2

 

 

Top image: Microsoft