What is artificial intelligence?

Artificial intelligence are computer systems that can perform tasks that typically require human intelligence, such as learning, reasoning, and problem-solving (e.g., voice assistants, autonomous driving systems, facial recognition, generative AI like ChatGPT).

What is NOT artificial intelligence?

Computer systems that strictly follow predefined, fixed rules without the ability to learn from data or adapt to new situations (e.g., basic calculators) are not considered artificial intelligence.

How are Large Language Models (LLMs) related to the broader fields of deep learning, machine learning, and AI?

Large Language Models (LLMs) are a type of deep learning model, which is a subset of machine learning, all of which fall under the broader umbrella of Artificial Intelligence (AI). LLMs utilise neural networks with many layers to process and generate human-like text, representing a significant advancement in AI's language capabilities.

What have been recent advancements in AI?

AI systems routinely exceed human performance on standard benchmarks, such as image classification (2015), basic reading comprehension (2017), visual reasoning (2020), and natural language inference (2021). Large Language Models (LLMs) have become multimodal: they can generate fluent text in dozens of languages, process and generate audio, and create and explain visuals, including 3D.

What are AI’s most significant challenges?

AI cannot (yet) reliably deal with facts, perform complex reasoning, explain its conclusions, or react appropriately to unforeseen circumstances.

What are the two futures of ai?

Scenario 1: Technology continues to improve and is increasingly used, having major consequences for productivity and employment.
Scenario 2: Adoption of AI remains constrained by the current limitations of the technology and the availability of quality data (AI Index 2024).

What tangible benefits has AI brought for non-AI companies to date?

AI adoption among corporates has surged from 50% to 72% over the past six years, with increased usage across all regions and industries, particularly in professional services. Companies are now utilising AI in more areas of their business, with half of the respondents to a McKinsey survey reporting adoption in two or more functions. Investments in both generative and analytical AI are yielding results. For example, generative AI is generating cost savings in HR (e.g., automating candidate screening and onboarding processes), while analytical AI is driving revenue increases in supply chain and inventory management (e.g., through predictive analytics and demand forecasting). Additionally, analytical AI continues to drive cost savings in service operations and revenue growth in marketing and sales (McKinsey 2024).

What is the impact of AI on productivity?

Economist Daron Acemoglu (MIT) forecasts that generative AI will provide a 'nontrivial but modest' boost to US productivity and GDP growth over the next few decades, estimating increases of 0.7% and 1.1%, respectively. Goldman Sachs predicts that U.S. productivity growth could rise by 1.5% annually over the next decade. McKinsey projects that the overall impact of AI and other automation technologies could contribute to a 1.5 to 3.4 percentage point increase in average annual GDP growth in advanced economies during the same period (Acemoglu 2024).

What are potential risks of investing into the AI sector?

The AI sector is experiencing a surge in valuations (Crunchbase 2024), raising concerns about a potential bubble. Some traditional AI business models (e.g., chatbots, search engines) may be overvalued due to the hype surrounding the technology. Upcoming regulations and intellectual property disputes could further consolidate the market by increasing costs for smaller companies. The development of general-intelligence AI could lead to market consolidation, favouring established players and potentially leading to the defeat of specialised AI providers (Proskauer 2024).

Rising AI valuation premium

How to value AI companies?

Pre-revenue valuations often focus on the strength of the AI technology, team expertise, and market potential, using methods like the Venture Capital method and benchmarking. For post-revenue companies, traditional metrics come to the forefront. IP, regulations, and investor sentiment also impact valuations.