[{"data":1,"prerenderedAt":29},["ShallowReactive",2],{"nr-en-chatgpt-llm-bewerbungsauswahl-vorurteile-studie":3},{"slug":4,"title":5,"dek":6,"date":7,"time":8,"publishedAt":9,"updated":10,"updatedAt":10,"dateFmt":11,"updatedFmt":10,"kind":12,"tier":13,"author":14,"authorName":15,"topics":16,"tracker":22,"trackerLabel":23,"headlineStat":24,"image":25,"ogImage":25,"imageAlt":5,"csv":10,"minutes":26,"words":27,"html":28},"chatgpt-llm-bewerbungsauswahl-vorurteile-studie","Study: ChatGPT and LLMs Show Massive Bias in Recruitment Screening","Researchers at Princeton University warn: When companies use Large Language Models to pre-screen job applications, these systems not only amplify human prejudices – they develop new ones. This poses a significant risk for German HR departments.","2026-07-23","04:43","2026-07-23T04:43:00+02:00","","July 23, 2026","daten","standard","ideal-syka","Ideal Syka",[17,18,19,20,21],"AI in Recruitment","Algorithmic Bias","HR Technology","Discrimination","LLMs","\u002Feu-ai-act-fahrplan","EU AI Act: High-Risk AI in Recruitment","LLMs stereotype job applicants more strongly than humans","\u002Fog-nr\u002Fchatgpt-llm-bewerbungsauswahl-vorurteile-studie.en.png",2,430,"\u003Cp>Your CV could soon be evaluated by an AI before any human ever sees it. Many companies are already using Large Language Models like ChatGPT to automate the pre-screening of job applications. But a new study from \u003Cstrong>Princeton University\u003C\u002Fstrong> now shows: These systems develop severe biases when performing this task – and stereotype candidates more strongly than humans would.\u003C\u002Fp>\n\u003Ch2>Key Facts\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Researchers at Princeton University\u003C\u002Fstrong> investigated how LLMs behave in job application screening\u003C\u002Fli>\n\u003Cli>LLMs not only adopt human biases from training data, but \u003Cstrong>develop additional bias through their work experience\u003C\u002Fstrong>\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Agentic AI systems\u003C\u002Fstrong> (those that remember things and learn from them) particularly amplify this effect\u003C\u002Fli>\n\u003Cli>The systems stereotype applicants more strongly than human recruiters\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>How AI Creates Bias on Its Own\u003C\u002Fh2>\n\u003Cp>Scientists have long known that LLMs absorb human biases from their training data. However, the new study points to an additional problem: Models develop biases \u003Cstrong>from the experiences they gain during daily work\u003C\u002Fstrong>. This is particularly critical because agentic AI systems – those that remember things and attempt to learn from them – entrench these patterns and incorporate them into their decisions.\u003C\u002Fp>\n\u003Cp>This means: The longer an LLM screens applications, the stronger its biases can develop and become embedded.\u003C\u002Fp>\n\u003Ch2>Why This Is Problematic for Companies\u003C\u002Fh2>\n\u003Cp>For German companies introducing AI-driven recruitment tools, this creates a dual risk. First, there are \u003Cstrong>legal consequences\u003C\u002Fstrong>: Employment law prohibits discrimination in hiring – and an automated system that systematically disadvantages certain groups violates this. Second, there is a \u003Cstrong>reputational risk\u003C\u002Fstrong>: If it becomes known that a company overlooks qualified candidates due to AI bias, it can cause significant damage.\u003C\u002Fp>\n\u003Cp>Additionally: Companies that miss top talent because an algorithm filters them out lose in the competition for skilled workers.\u003C\u002Fp>\n\u003Ch2>What Does This Mean in Practice?\u003C\u002Fh2>\n\u003Cp>This doesn't mean AI in recruitment is fundamentally wrong. But it means companies \u003Cstrong>cannot blindly trust automated pre-screening\u003C\u002Fstrong>. What's needed:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Regular audits\u003C\u002Fstrong> of AI systems for bias\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Human oversight\u003C\u002Fstrong> of system decisions\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Transparency\u003C\u002Fstrong> toward applicants about AI use\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Documentation\u003C\u002Fstrong> of decision criteria\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>The Princeton University study thus provides important evidence for what many HR professionals already suspect: AI can be a tool in recruitment, but only if properly monitored.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Implications for German companies:\u003C\u002Fstrong> These findings should alert HR departments. Companies using AI systems for application screening bear responsibility for their fairness – not just morally, but legally. Automation without oversight is a risk no company can afford.\u003C\u002Fp>\n\u003Ch2>Sources\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Ca href=\"https:\u002F\u002Ft3n.de\u002Fnews\u002Fstudie-zeigt-chatgpt-und-co-entwickeln-bei-der-bewerberauswahl-heftige-vorurteile-1753740\">t3n\u003C\u002Fa>\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cem>Editorially owned by \u003Ca href=\"\u002Fen\u002Fautor\u002Fideal-syka\">Ideal Syka\u003C\u002Fa>. Sources and method: \u003Ca href=\"\u002Fen\u002Fredaktion\">Newsroom & method\u003C\u002Fa>. Tips and corrections: \u003Ca href=\"mailto:ai@i6eal.de\">ai@i6eal.de\u003C\u002Fa>.\u003C\u002Fem>\u003C\u002Fp>\n",1784788189268]