The growth of AI in the workplace shows no signs of abating - and nor does the concern of businesses that they will be left behind in the AI arms race. At the same time, there is increasing awareness of the legal, practical and ethical risks AI can pose. Here are some of the key trends we expect to see dominating this discussion in 2025:
Agentic AI One of the criticisms levelled at ChatGPT and other generative AI applications is that they are reactive rather than proactive - in response to instructions (prompts) they "generate" material (text, in the case of LLMs, based on vast amounts of training data and predictive tools). The tools often produce impressive-looking results (although the problem of hallucinations is well-publicised). But agentic AI goes a step further, operating with a degree of autonomy so that it doesn't just complete tasks as instructed, but can identify what tasks need to be done and respond to a developing situation more readily. It's a more sophisticated and proactive application of AI technology, and one which is increasingly being adopted in a workplace context.
Sustainability As the consequences of climate change become increasingly apparent, expect there to be an increasing focus on the vast energy consumption associated with AI data centres, as well as the working conditions of the “ghost workers” , often low paid and precariously employed, who help to train AI models. Consumers, investors, campaigners and regulators are likely to scrutinise closely businesses which rapidly expand their AI adoption without consideration of the environmental and social issues surrounding it.
Data security AI is increasingly being used in sophisticated frauds and scams, with even CFOs being caught out. Businesses investing in AI will need to do their due diligence on the cybersecurity implications - but, at the same time, AI tools are likely to play a huge role in combating fraud. With the new criminal offence of failure to prevent fraud, and employee fraud a major risk issue for many organisations, we expect that businesses will be looking to invest in the anti-fraud tools now proliferating in the market.
Bias and discrimination Again, AI cuts both ways here. The issue of biased outcomes from AI tools is well-documented - but, at the same time, many AI tools promise to contribute positively to diversity, for example, by making job adverts and interview questions more accessible to neurodiverse candidates. We expect to see these claims come under increasing scrutiny in the year ahead, as a tough economic climate forces businesses to invest only in technology which can prove its worth.
I recently discussed some of these issues on the Meritas Employment podcast - have a listen here.
Want to hear more from us? 
