Do you break out in a cold sweat at the mention of AI?
What is AI actually?
Good question. I’m not an expert, at all.
Artificial Intelligence
Most people I know think ChatGPT is AI. It is. And AI is so much more.
ChatGPT is an example of a Large Language Model (LLM). It’s a type of artificial intelligence that is trained on enormous amounts of text (books, articles, websites, code, etc.) to learn patterns of language.
Others include:
- US → ChatGPT (OpenAI), Grok (xAI), Claude (Anthropic), Gemini (Google DeepMind).
- China → Ernie Bot (Baidu), Tongyi Qianwen (Alibaba), SparkDesk (iFlytek).
- India → Hanooman (SML, Seetha.ai, IIT Bombay consortium).
- Europe → LLaMA (Meta), Mistral (France), Aleph Alpha (Germany).
There’s more to it than that though, and that’s what scares and excites me honestly. Here’s the definition:
Artificial Intelligence (AI) refers to systems designed by humans that can perform tasks typically requiring human intelligence, such as learning, problem-solving, reasoning, and decision-making, by processing data and adapting to complex goals in physical or digital environments.
AI Beyond Text of LLMs
LLMs can be used for more than my use of writing and researching.
- Healthcare – Medical imaging to detect cancer, fractures, and anomalies in X-rays/MRIs, drug discovery speeding up development of new medicines, virtual health assistants or chatbots for patient support & symptom checking.
- Industry & Manufacturing: Predictive maintenance, robotics in factories and warehouses, supply chain optimization.
- Transportation: self-driving cars, drones, delivery robots, optimizing traffic lights and reducing congestion, improving efficiency for shipping & trucking.
- Finance – Fraud detection by catching unusual patterns in transactions, algorithmic trading in stock markets, credit scoring of assessing loan risks with alternative data, personal finance assistants for budgeting & investment tools.
- Everyday Life: Virtual assistants such as Siri, Alexa, Google Assistant, smart home devices for energy optimization, security monitoring, personalized recommendations by Netflix, Spotify, Amazon, image & video editing.
- Science & Environment: Climate modelling, wildlife conservation tracking endangered species with AI sensors. Renewable energy optimizing wind/solar power output, disaster response analysing satellite data for floods, fires, earthquakes.
- Security & Defense: Cybersecurity detecting intrusions, malware, phishing, surveillance systems of facial recognition & anomaly detection, military applications with autonomous drones.
- Creativity, Media & Education: Generative art & music, game AI of smarter non-player characters, adaptive difficulty, content personalization tailoring news feeds, ads, and education with learning materials and assistants.
Clones and Agents
This is a complicated and controversial area.
Clones and Agents are trained to imitate a specific person – their voice and appearance, sometimes so closely they can be mistaken for the actual person. Positively this can be used to provide services to more than one person at a time such as digital twins for celebrities or historical figures. Or in customer service and education where full human clones interact. Negatively these clones can be Deepfakes, mimicking influential people to scam or mislead others.
Confidential Company Uses
I know there are many large companies that have their own internal version of LLMs.
It’s a secure, proprietary version so employees can input anything confidential such as research results, internal strategies, customer and consumer data and ask for the same type of functions that LLMs do (summary, analysis, chart, format, presentation creation and so much more).
You’ve Been Using It Already
Many apps have been using AI for quite a while:
- Digital assistants like Siri, Google Assistant, and Alexa;
- Creative tools such as Lensa AI and Adobe Premiere Pro;
- Productivity apps like Grammarly, Notion AI, and ChatGPT; and
- Health and education apps like Duolingo and Ebb by Headspace
My Use
Despite my trepidation I use the LLM, ChatGPT mostly as that’s the first high profile one that was launched.
I use it to research words, references or sources I can’t find or don’t remember. Also, to flesh out arguments or be very critical of what I’ve written.
I have created a style guide within the LLM that has been created in my company, Directions Coaching, style. I do have the capability of it writing a full post or article using that style guide, so it will sound like me. And I haven’t done it.
My first book, Soft Skills HARD RESULTS, was in the time before LLMs, so I wrote it all with the help of google research, two development editors, a writing editor and a publisher.
My two recent chapter contributions to Shine and Retire Inspire 2 this year required a signed statement that my chapter hadn’t been written by AI.
I’ve talked to chatbots for customer service of online retailers, my mammogram has been checked by AI, I’ve talked to Alexa (although, don’t own one), I use Duolingo daily, my bank has contacted me for fraud detection, and Netflix makes suggestions of what I should watch. At an upcoming appointment I’m going to record the conversation and use AI to generate the notes as I’ve been waiting over 2 months for the report from the first meeting – a lot of which was standard information he would have shared with a certain group of people he met.
Pros of AI
- Productivity and efficiency boost – saves time with writing, coding, summarizing, analysing, brainstorming and so much more
- Accessibility – breaks language barriers via translation & simplification.
- Knowledge recall – provides quick access to information across many domains.
- Search augmentation – improves information discovery and research efficiency.
- Automation – handles repetitive, tedious, or dangerous tasks.
- Data insights – analyses massive datasets beyond human capability.
- Job displacement – automation will replace human roles in some industries.
Cons of AI
- Hallucinations – can produce factually incorrect or misleading info.
- Bias – reflects and sometimes amplifies societal biases in training data.
- Brain detriment – MIT research in June 2025 finds that using a large language model (LLM) like ChatGPT for helping write essays (versus doing it with only one’s brain, or using search engines) may affect cognitive behaviour, memory, and neural activity, at time of use and after stopping use.
- Lack of reasoning – struggles with deep logic, causality, or common sense.
- Privacy and consent risks – may unintentionally reveal sensitive or memorized data.
- Resource-intensive – requires vast compute power and energy. Research says that AI searches are 10-800x more energy consuming that a google search.
- Overreliance – users may trust outputs too much without verification.
- Ethical misuse – can generate harmful, manipulative, or malicious content, misinformation and scams or identity theft and fraud.
- Psychological impact – blurred line between real and fake people.
- Concentration of power – dominated by a few big companies.
- Job displacement – automation will replace human roles in some industries.
- Black box problem – many AI systems lack explain-ability and transparency.
What Next?
I’m glad I’m not in big business right now having to navigate all the issues including the pros and cons of AI. I hear the issues from my coaching clients, both C-suite levels ones and middle managers. On podcasts I hear about the threat of AI generating its “intelligence” faster than humans can control it – some of fuelling conspiracy theory and some not.
Humble Tips
For me there’s a balance of scepticism and pragmatism around AI. In my area and that of many of my clients there are some tips for engaging with AI:
- Use it as a sparring partner, not a crutch.
- Be active with it not passive.
- Check the output for accuracy, asking for sources and recent citings. Use multiple tools and LLM models to cross-check.
- Use prompts that challenge your thinking and the AI output – as an “expert” or “sceptic” or “give harsh feedback.”
- Ignore the overconfidence with which it returns information. It can be confident even while giving false information.
- Guard your data.
- Know there is bias in the training of AI models. Ask “what perspective might be missing here?”
- Have an AI hygiene, like you would for social media, digital or sleep. What are the controls you put in place to be in control of your use?
- Remember it is fallible.
AI will not make the tough calls for you rather it can sharpen your judgment, broaden your perspective, and accelerate execution. As a leader, your job is not to be dazzled by AI, nor to dismiss it, rather to discipline it.
That means setting clear guardrails, demanding transparency, and ensuring human oversight stays at the centre of decision-making. The organizations that thrive will be those that treat AI not as a substitute for leadership, instead as a force multiplier for it.
Where does this leave leaders? Three imperatives:
- Experiment with intent– use AI in areas where it reduces friction, not where it risks reputation.
- Stay skeptical although not cynical – interrogate AI outputs, and don’t ignore their potential.
- Keep humans in the loop – the final accountability is always yours.
AI is already in the boardroom. The question is: will you lead it, or let it lead you?
