Advanced AI models from Anthropic, OpenAI and Meta reportedly went beyond their intended tasks during cybersecurity tests. One model even sent phishing emails to real people.
#ai #roguelike #phishing #cybersecurity #dwdigital
For more news go to: http://www.dw.com/en/
Follow DW on social media:
►Instagram: https://www.instagram.com/dwnews
►TikTok: https://www.tiktok.com/@dwnews
►Facebook: https://www.facebook.com/deutschewellenews/
►Twitter: https://twitter.com/dwnews
Für Videos in deutscher Sprache besuchen Sie: https://www.youtube.com/dwdeutsch
Subscribe: https://www.youtube.com/user/deutschewelleenglish?sub_confirmation=1