As AI adoption grows, so too does the practical AI confidence gap facing unemployed people and those trying to get back into work. There is a growing need to equip people with the skills needed to use AI responsibly day to day and the confidence to put these into practice and make judgements on using AI safely at home and in work.
Many people report feeling unsure about the extent to which AI can be trusted, used to inform their work or relied upon for day-to-day tasks. This confidence gap is a real barrier to getting people into work. People need to be equipped with the knowledge and confidence to make their own decisions and be empowered to use their own judgement.
Responsible AI use is not just about rules and governance It is about education, and supporting people with the confidence and skill to use AI as a real enabler for their job search and into a new role, not as a replacement of their own skills or in a way that puts them at risk.
We know that learners find training useful: among learners who reported low confidence of less than 5 out of 10, average confidence grew from 2.5 to 8.0 by the end of their training, and for 97% of learners, training improved their job search. For people looking to enter the new, AI powered world of work, having the know-how on responsible AI use and the judgement to apply those skills is invaluable.
The need for responsible AI
The potential benefits of AI are no secret to anyone. When used in the right way, AI tools can be invaluable enablers, transforming efficiency and unlocking new opportunities. Through practical AI upskilling, people are equipped with the skill to navigate this new world safely and in a way that can best support them. People searching for employment can benefit from these skills and take them into a new role.
A report from McKinsey found that demand for AI fluency skills has increased nearly sevenfold between 2023 and 2025, indicating AI-related skills are now a requirement and not just a nice to have.
But AI doesn’t come without risks. The National Cyber Security Centre publishes guidance on these tools, stating that LLMs such as ChatGPT are only as good as the data they are trained on, so need to be treated with caution. They can also contain some serious flaws in the form of hallucinations, support biases, or manipulate the data used to train it.
There is also the threat of shadow AI: the use of AI tools without the approval, knowledge or oversight of IT and security teams. For those both in and out of work, it is important to understand how to use AI securely, and where it can bleed into becoming a risk.
Feeding sensitive information into public facing LLMs for example can leave businesses and people vulnerable to leaks and lead to unauthorised data access, regulatory failures or security breaches.
This is exactly why responsible AI training is important to keep people safe and maintain trust in LLM tools.
Responsible AI in practice
Strict policies are often relied upon as guidance for AI usage, but rigid rules and principles often come unstuck in the real world when people face ambiguous scenarios that don’t quite fit into them neatly. They also don’t work to identify and tackle the gaps in confidence that people have, as can be done with the right training.
Responsible AI starts with people not technology, and people rarely respond to rules without understanding the reasoning behind them. AI governance frameworks alone are not enough without training and real-world understanding. All the tools and programmes in the world can be at someone’s fingertips but without confidence, training and guidance, there’s little to no chance of getting the most out of them.
For people entering work, using AI tools like LLMs can often be intimidating. Good training should focus on practical decision making in AI use, looking at how to identify bias in data and evaluate the veracity of sources. Being able to evaluate model outputs for reliability is crucial, as is knowing when to incorporate human review more closely and how to approach suspicions of misinformation.
This isn’t a one-off tick-box exercise for people, but an in-depth look at practical examples when using AI, in scenarios that are relevant to the job. People should leave with the confidence to question assumptions in LLMs, navigate and test models to suit them and be able to apply the information in real world situations.
AI skills are people skills
The key skills people need to use AI responsibly are not as technical as you might think and do not require any in-depth understanding of how to build an LLMs, rather they are forms of judgement. These include critical thinking, risk awareness, prompt literacy, and human oversight.
In many cases, these are extensions of skills we already use in our day-to-day lives. Training should cultivate habits of curiosity, accountability, and structured scepticism. People need confidence at work to challenge decisions, raise their concerns and dig further into the data.
There are also model specific skills that are becoming more commonplace in the world of work. Prompt design, meaning knowing how to word requests to get the best results for your task, is of growing importance, as is the ability to structure tasks so models are used in the right contexts and with the right safeguards.
Responsible AI also depends on people developing the practical skills to work effectively alongside AI systems, rather than simply using them passively. People need to know when AI is appropriate, when human expertise is essential, and how to combine both effectively.
Confident AI usage combines technical awareness with good judgement, balancing reliance on tools with practical skills.
Success is confident, practical AI usage
AI is a crucial business enabler but is only effective when people across an organisation or seeking to join it have the skills and confidence to use it in practice. AI literacy, critical thinking, risk awareness, and human judgment are quickly becoming foundational workplace skills.
The success of AI will be shaped by the investment of time in people to build the practical skills needed to use AI tools safely and accountably, to narrow the confidence gap facing people looking to get into work.
Training on responsible AI for people both in and out of the workplace must address this gap head on, with into co-ordinated, targeted AI training that supports people with the basic skills to use AI safely and securely.
