What are peoples thoughts on 'AI coding platform's flaws allow people to be hacked'?

Hi, I was genuinely excited to discover Playwright along with Planner, Generator, and Healer. After 17 years in manual testing, tools like these would be a huge step forward for me. Unfortunately, my organisation declined the installation based on this BBC article: https://www.bbc.co.uk/news/articles/cy4wnw04e8wo
I’m interested in people’s views on the headline “AI coding platform’s flaws allow people to be hacked.”
Is there any evidence that these concerns apply specifically to Microsoft’s Playwright?