Regular readers of our blogs will know that we’ve written extensively about the growing role of artificial intelligence (AI), not just within marketing, but across hiringtoo.
However, the growth of these platforms has led to concerns in some circles around potential biases; with this in mind how can employers ethically use AI in marcomms recruitment?
AI in marcomms recruitment
We should start by saying that AI is already being used as a key part of many hiring programmes, particularly at larger firms, and allows employers to identify candidates and review their suitability for roles far quicker than they were able to in the past, amongst other benefits. When these tools first emerged, they were initially mooted as having the potential to remove biases from the recruitment process. In fact, they’re likely having the opposite effect. As with any transformative technology, adoption comes with uncertainty around ethical responsibilities. For those seeking marcomms expertise, the question isn’t just about leveraging AI but rather doing so in a way that upholds fairness, inclusivity, and human-centred values.
Several of the leading GenAI tools based on large language models (LLMs) have already been under the spotlight for the potential biases they generate in many user answers. The so-called ‘knowledge’ of these platforms is built on a foundation of pre-existing data and, ultimately, the attitudes of the individuals or groups feeding it this information. In practice, this means that potential biases are an inherent part of the platforms, and hard to avoid without careful management. But not doing so can perpetuate discrimination related to gender, race, age or disabilities, amongst other areas.
Transparency lacking
The likes of ChatGPT, Microsoft’s CoPilot and DALL-E also face problems with a lack of transparency and understanding as to how they operate at a fundamental level. Put simply, few people actually understand why these platforms work in the way they do, and how they learn and take on new knowledge. While it's still
