Unlike previous developments, these text-to-image tools quickly found their way from research labs to mainstream culture<\/a><\/strong>, leading to viral phenomena such as the \u201cMagic Avatar\u201d feature in the Lensa AI app, which creates stylised images of its users.<\/p>\n In December, a chatbot called ChatGPT stunned users with its writing skills<\/strong><\/a>, leading to predictions the technology will soon be able to pass professional exams<\/a><\/strong>.<\/p>\n ChatGPT reportedly gained one million users in less than a week. Some school officials have already banned it<\/a><\/strong> for fear students would use it to write essays. Microsoft is reportedly<\/a><\/strong> planning to incorporate ChatGPT into its Bing web search and Office products later this year.<\/p>\n What does the unrelenting progress in AI mean for the near future? And is AI likely to threaten certain jobs in the following years?<\/p>\n Despite these impressive recent AI achievements, we need to recognise there are still significant limitations to what AI systems can do.<\/p>\n Recent advances in AI rely predominantly on machine learning algorithms that discern complex patterns and relationships from vast amounts of data. This training is then used for tasks like prediction and data generation.<\/p>\n The development of current AI technology relies on optimising predictive power, even if the goal is to generate new output.<\/p>\n For example, GPT-3, the language model behind ChatGPT, was trained to predict what follows a piece of text. GPT-3 then leverages this predictive ability to continue an input text given by the user.<\/p>\n \u201cGenerative AIs\u201d such as ChatGPT and DALL-E 2 have sparked much debate<\/strong><\/a> about whether AI can be genuinely creative and even rival humans in this regard. However, human creativity draws not only on past data but also on experimentation and the full range of human experience.<\/p>\n Many important problems require predicting the effects of our actions in complex, uncertain, and constantly changing environments. By doing this, we can choose the sequence of actions most likely to achieve our goals.<\/p>\n But algorithms cannot learn<\/a><\/strong> about causes and effects from data alone. Purely data-driven machine learning can only find correlations.<\/p>\n To understand why this is a problem for AI, we can contrast the problems of diagnosing a medical condition versus choosing a treatment.<\/p>\n Machine learning models are often helpful for finding abnormalities in medical images \u2013 this is a pattern recognition problem. We don\u2019t need to worry about causality because abnormalities are already either present or not.<\/p>\n But choosing the best treatment for a diagnosis is a fundamentally different problem. Here, the goal is to influence the outcome, not just recognise a pattern. To determine the effectiveness of a treatment, medical researchers run randomised controlled trials. This way, they can try to control any factors that might affect the treatment.<\/p>\n The confusion between these two types of problems sometimes leads to suboptimal applications<\/a><\/strong> of machine learning in organisations.<\/p>\n While the success of recent work in AI demonstrates the value of data-driven models, many of the problems we would want computers to solve require an understanding of causation.<\/p>\n Current AI systems lack this ability, apart from specialised applications such as board games<\/a><\/strong>.<\/p>\n A Stable Diffusion artwork generated by the prompt \u2018The limits of artificial intelligence\u2019. Current AI technology tends to perform poorly in unexpected situations. In this case, the prompt is not well represented in Stable Diffusion\u2019s training data. [This article’s thumbnail image was generated in Midjourney<\/a><\/strong> using the same prompt.]<\/p><\/div>\n Language models such GPT-3 and ChatGPT can successfully solve some tasks requiring common-sense reasoning.<\/p>\n However, the following interaction with ChatGPT, adapted from an experiment by Gary Marcus<\/a><\/strong> suggests it is not entirely reliable in this respect.<\/p>\n I\u2019m eight years old. When I was born, my biological mother was in Barcelona and my father was in Tokyo. Where was I born? Think through this step by step.<\/strong><\/p>\n It is not mentioned where you were born. All we know is that your biological mother was in Barcelona and your father was in Tokyo at the time of your birth.<\/p><\/blockquote>\n Whether AI systems such as ChatGPT can achieve common sense is a subject of lively debate among experts.<\/p>\n Sceptics such as Marcus point out we cannot trust language models to robustly display common sense since they neither have it built into them nor are directly optimised for it. Optimists argue that while current systems are imperfect, common sense may spontaneously emerge<\/a><\/strong> in sufficiently advanced language models.<\/p>\n Whenever groundbreaking AI systems are released, news articles and social media posts documenting racist<\/a><\/strong>, sexist<\/a><\/strong>, and other types of biased<\/a><\/strong> and harmful behaviours<\/a><\/strong> inevitably follow.<\/p>\n This flaw is inherent to current AI systems, which are bound to be a reflection of their data. Human values such as truth and fairness are not fundamentally built into the algorithms \u2014 that\u2019s something researchers don\u2019t yet know how to do.<\/p>\n Let us invent then a new breed of AI systems that mix an awareness of the past with values that represent the future that we aspire to.<\/p>\n Our focus should be on figuring on how to build AI that can represent and reason about *values*, rather than simply perpetuating past data.<\/p>\n \u2014 Gary Marcus (@GaryMarcus) April 19, 2021<\/a><\/p><\/blockquote>\nAI excels at pattern recognition<\/h2>\n
Cause and effect<\/h2>\n
![\"\"](\"https:\/\/mybroadband.co.za\/news\/wp-content\/uploads\/2023\/01\/Stable-Diffusion-limits-of-artificial-intelligence.jpg\")
<\/a>Common sense reasoning<\/h2>\n
Human values<\/h2>\n
\n