Can AI help employees to upskill?

A recent MIT study suggested that the use of artificial intelligence (AI) may harm critical thinking. It was not the first study that came to that conclusion: A Microsoft poll of knowledge workers earlier this year also implied that GenAI usage may have a negative impact on thinking abilities. It argued that GenAI shifts the nature of critical thinking towards information verification, response integration and task stewardship, and observed that when users had higher confidence in GenAI, this appeared to be associated with less critical thinking. 

While these studies make great headlines, the use of AI tools has, however, become an undeniable reality in everyday business. AI usage will only grow. And experience shows us that rather than making us less intelligent, AI can be a powerful tool in helping employees upskill. The right use of AI can make workers smarter, not less capable.

The study findings resonate because many users who are new to AI tools tend to hand tasks to AI without thinking. Asking ChatGPT or Microsoft Copilot to produce entire reports and presentations means previously time-consuming jobs can now be completed with minimal effort. This ‘do it for me’ mindset creates the illusion of efficiency, but it reduces the learning opportunity. And without human input, the results may look shiny but often turn out to be flawed.

But, when we accuse AI of dumbing us down, perhaps we are blaming the wrong thing. AI isn’t the problem. The way people use it is.

Stop using AI as a shortcut

The behaviour that can damage critical thinking is misusing AI as a shortcut to complete whole tasks – thereby avoiding work entirely. This over reliance is a common problem. Take the example of writing a report. An employee might share bullet points, documents or a call recording with ChatGPT and tell it to create the full report from scratch. This cuts the time spent on the report from 8 hours to 3 minutes – and the employee’s own input to zero, with a very likely impact on the quality of the report.

There is another, better way. Instead of asking AI to think for them, employees should think with the technology. Just like an assistant, an AI tool can provide support. It can help workers develop and augment their skills – not to mention do much of the legwork, but it still needs human oversight. A middle ground for producing the report would be to let AI collate, tidy up and structure information and draft an outline. The human’s job then is to take over and edit, expand, verify and refine the draft.

When used in this way, the time spent on the report might drop from 8 hours to 2 hours rather than minutes. This is still a considerable time saving, and it gives a better result. And with this more intelligent and effective way to use AI, learning improves, too.

AI is already a teacher

Many employees are already developing skills and absorbing knowledge through AI every day without realising it. Even in the most basic application of AI, where it essentially serves as a smarter search tool, the way workers use it has quietly shifted from “searching for information” to “interacting with a tutor”. This is because AI doesn’t only deliver the facts, but can also provide more detailed insights, examples, comparisons and step-by-step guidance, when asked. The technology is fantastic at researching, sifting through information and highlighting patterns, all of which help users digest information faster.

Increasingly, users are also harnessing AI tools to teach them new knowledge. Examples for this might be understanding product management frameworks or learning new programming languages. These are hybrid tasks where users don’t just ask AI to do something for them but also learn through the interaction – for instance, by requesting practical scenarios or explanations at different levels of complexity. All of which helps them develop new skills at their own pace.

This shift is healthy, as long as organisations guide it. Indeed, when used well, AI could become one of the most powerful upskilling accelerators the workforce has ever had – helping employees learn faster, be better informed, think more broadly and ultimately, perform at a higher level.

AI can be transformative

Of course, there are limits to what AI can teach. Due to its nature, it is best for knowledge that can be acquired in self-directed study – such as how to build a survey questionnaire or structure a CV – and that don’t require formal training or specialist qualifications. Because it can sort through facts fast and effectively summarises the most common expert answers from across the internet, it teaches best practices, helping even junior colleagues punch above their weight.

One good example for upskilling through AI is in product management. To determine their strategy, product managers need to first understand markets, conduct competitive analysis and gather sentiment from real users. AI is extremely good at this because it can process huge volumes of online content from sources such as Reddit, reviews and forums and provide an almost immediate snapshot of customer sentiment, as well as pinpointing strengths and weaknesses. Armed with this knowledge, the employee can work at a higher level and much faster.

On the other hand, while AI can support the learning process, it is less powerful when it comes to teaching highly advanced disciplines that require supervised practice and formal accreditation.

Another limitation is that AI delivers information but doesn’t automatically test you. Many users don’t realise that they can request quizzes, exercises or assessments to consolidate their learning.

Adoption isn’t optional

In three years’ time, nearly every organisation will use AI for employee skill development, whether they plan to or not. Trying to ban AI would be unrealistic. The most forward-thinking companies are already buying tools and encouraging AI-supported learning. While currently, this mainly happens through general-purpose LLMs, in the future, we will see specialised AI learning tools, AI powered courses and guided learning experiences, as well as more audio and video-based output formats.

However, AI will supplement, not replace traditional training – just like online learning didn’t replace textbooks. People have different learning preferences and will expect a mix of materials, interactive learning and offline courses. Additionally, while it’s time to rethink how we teach skills, formal evaluation and accreditation still matter.

For those organisations wanting to embrace the trend now, the advice is to lead responsibly. Many of the day-to-day capability and productivity gains now happen inside AI tools, but AI usage needs guardrails. As well as providing company-approved, secure AI platforms, businesses should establish and enforce strong data protection and privacy guidelines. They should stipulate that AI tools need human oversight. Training is key, too, teaching employees how to use AI well, encouraging critical thinking and setting expectations around source verification and careful fact checking instead of over-reliance on AI outputs.

Organisations should also foster a culture of healthy scepticism. AI can mislead when it relies on the wrong data. But its strength is that it lets users interrogate the information presented. Learners should always be encouraged to ask for evidence, cross-check and iterate.

Critical thinking isn’t harmed by AI. It is harmed by unquestioning use. When embraced with the right mindset, it can make us smarter. Organisations that lean into AI for upskilling will build stronger workforces. Those that resist will fall behind.