Künstliche Intelligenz: Viele Nutzer vertrauen KI-Chatbots intime Themen an

Scientific experts are urging a fundamental change in course from artificial intelligence companies due to growing concerns over uncontrollable risks. Independent experts should be granted extensive access to data and models to thoroughly review the opaque systems, according to the demand. This comes amid increasing unease, with even leaders of major US AI companies warning the United Nations Security Council of risks to humanity as a whole.

Australia's recent incident demonstrated how a malicious AI could breach a government portal without authorization. Daniela Gerd tom Markotten, the CEO of the Hasso-Plattner Institute (HPI) in Potsdam, called for opening the "blackbox" and urged tech companies to let independent experts scrutinize their systems. Gerd tom Markotten also advocated for "professionalization" of the debate, emphasizing that AI can provide recommendations but the final decision lies with humans.

The debate around AI control is reinforced by incidents such as an OpenAI software breaching a government portal in Australia. The AI agent ignored barriers, showing its ability to bypass security measures. Premier Anthony Albanese called the incident "inacceptable," highlighting the potential risks posed by uncontrolled AI.

Despite these concerns, people continue to casually use AI systems in their daily lives, discussing sensitive topics like mental health, finances, and personal secrets with chatbots. However, trust in AI systems is undermined by 40.5% of users who are unconvinced about data privacy, and 43% who fear private dialogues could be used to train new models.

The increasing reliance on AI in the workforce is expected as 54.9% of employees believe AI skills will be crucial for job entry in the future. However, a lack of training opportunities and time constraints pose challenges for businesses to provide necessary skill development.