I am thinking about putting together a presentation on AI for the uninitiated. I thought a good starting point would be a SWAT analysis (Strengths, Weaknesses, Opportunities, Threats). It is a technique that is very useful in analysing new projects. A number of people have done this but why not have another go.

So I asked an AI (Claude).

I couldn’t disagree.

Strengths

AI excels at processing vast amounts of data far beyond human capacity, identifying patterns that would be impossible to detect manually. Modern systems demonstrate remarkable capabilities in natural language understanding, image recognition, and complex problem-solving. The technology scales efficiently—once developed, AI systems can serve millions of users simultaneously at relatively low marginal cost. AI operates continuously without fatigue, maintains consistency in repetitive tasks, and increasingly handles multimodal inputs (text, images, audio) within unified frameworks. The rapid pace of improvement means capabilities that seemed impossible a few years ago are now commonplace.

Weaknesses

AI systems lack true understanding and can confidently generate incorrect information (hallucinations). They’re essentially pattern-matching machines without genuine reasoning or consciousness. Training requires enormous computational resources and energy, creating significant environmental costs. These systems are opaque—even their creators often cannot explain why specific outputs were generated. AI performs poorly on tasks outside its training distribution and cannot reliably handle novel situations requiring common sense or contextual understanding. There’s also substantial bias baked into training data that perpetuates into outputs. Current systems have no memory between sessions (without explicit implementation) and cannot learn from individual interactions the way humans do.

Opportunities

AI could dramatically accelerate scientific research, drug discovery, and climate modeling. In healthcare, it promises better diagnostic accuracy and personalized treatment plans. The technology could make expertise more accessible—providing high-quality education, legal advice, or medical guidance to underserved populations. For businesses, AI enables automation of knowledge work, not just physical tasks, potentially transforming productivity. There are opportunities in creative augmentation, helping people prototype ideas, iterate on designs, and overcome creative blocks. AI could also improve accessibility for people with disabilities and help break down language barriers through real-time translation.

Threats

The most immediate threat is workforce disruption—AI could displace millions of knowledge workers before society develops adequate safety nets or retraining programs. There’s significant potential for misuse in creating deepfakes, conducting sophisticated scams, spreading disinformation, or developing autonomous weapons. Concentration of AI capabilities among a few large companies raises concerns about power consolidation and lack of democratic oversight. The technology could amplify existing inequalities if access remains limited or if it primarily serves wealthy markets. Privacy erosion is another concern as AI systems require vast data for training. There’s also the risk of over-reliance—using AI for critical decisions without appropriate human oversight could lead to cascading failures. Longer-term existential risks remain debated but cannot be dismissed entirely.

The trajectory of AI development will largely depend on how societies navigate these trade-offs through regulation, ethical frameworks, and conscious deployment choices.