Establishing AI Threat Baselines: An Experimental Assessment of GPT-2.0 as a Phishing Attack Vector Among Digitally Literate University Students