Very often we limit ourselves to thinking about the direct impacts of an action or of an event. Yet, impacts, most of the time, go on. We face cascading effects. For an event, a decision, we thus should consider first the direct impacts, then the second-order impacts (the direct impacts of the direct impacts), then the third-order impacts… until the nth-order impacts.
An EU’s pro-Taiwan stance within the context of Europe’s need for critical commodities.
Very rapid (less than 6 months) adoption of a generalised use of generative IA across companies, NGOs, governments, administrations and public services in the Western world. 我们在下面粘贴了我们收到的答案 这第三个例子，有评论，考虑到广泛的影响。人工智能给我们的结果只是级联影响的元素。一个适当的、真正的战略远见和早期凋零工作应该建立在一个充分发展的模型之上。
Nota: This is an experiment with varying results. Some results given by the AI were excellent, others were less so. For example, it seems that the AI provided best results first. Then, if one asked the same question again, the results obtained did not improve massively or could even be less interesting. Comments and discussions on the results you obtained are welcome!
EU’s pro-Taiwan stance: Improved European access to Taiwanese critical commodities: Taiwan is a key producer of critical commodities such as semiconductors and rare earth elements. The EU’s pro-Taiwan stance will improve access to these commodities, which are essential for European defense capabilities.
美国违约： Internationally, there would be a shift away from the US Dollar as the world’s reserve currency.
Current GPT-based apps break in terms of logic, consistency and ability to combine factors on different levels starting from the 4th-order effects. This appears to us as logical because they are “only” based on AI language models. As a result, even though GPT-based apps may appear able to reason, actually they cannot really provide analysis, which requires subjective reasoning. Nonetheless, as they are, GPT-based apps can already yield interesting results, which may become tools – to be checked and verified – for human beings.
The GPT4 model has the potential to offer increased capabilities and advancements which may surprise us. The future will tell…
Increased productivity: AI could boost productivity in various sectors, leading to an increase in output and efficiency. Companies that adopt AI quickly could have a competitive advantage over those that don’t.
Hence to obtain the best result we would need to mix the answers obtained with test 1 and test 2, added to our own specific knowledge and understanding. AI, as often in our case, is a tool that can help us and that we can usefully integrate within our work, but not a definitive panacea… at least not yet when using GPT-3.5.
However, as far as our IA is concerned, we observe some difficulties to combine different second-order effects: for example, if in the Western world – those countries that adopt rapidly generative AI – the massive job displacement especially of better paid jobs leads both to unrest and to economic downturn, then will the countries having massively adopted AI still have a geopolitical advantage? Further, is it the very way we conceive the international society and the world order that could change?
Again, we are at too general a level to be able to truly use the results obtained to steer policy, yet, what our AI Pithia highlighted is definitely worth considering and incorporating in our analyses.
GPT 4 (once it becomes widely available) could be better at this task, or/and we need to improve how we handle the AI model to get better and more specific answers. Alternatively, LLMs – for now – may just not be able to properly handle such tasks and we would need different types of models.
自2022年3月爆发以来，乌克兰战争的地缘政治和经济影响与气候变化的连带后果相结合（Jean-Michel Valantin，"乌克兰战争、美国大旱和即将到来的全球粮食危机", The Red Team Analysis Society，2022年5月1日）。在2023年，这种地缘政治与气候的关系很可能会急剧恶化。这源于很可能是一个涡轮增压的厄尔尼诺事件的出现，以及乌克兰战争对全球农业的破坏性影响。
厄尔尼诺现象是周期性的气候-海洋事件，当赤道太平洋表面升温1至3年时发生。经常性地，这些事件在世界各地引发一连串的极端气候事件。之前的2016年的厄尔尼诺现象在历史上是很强大的。2023年的厄尔尼诺现象很可能更加危险，因为气候变化的迅速恶化将非常可能加强厄尔尼诺现象的强度（Paloma Trascasa-Castro, "2023年厄尔尼诺现象回归的四个可能的后果", The Conversation，2023年1月26日）。
During an El Niño event, the equatorial Pacific Ocean may heat up to 3°, thus heating up the whole atmosphere. The singular aspect of the 2023-2024 El Niño is that it happens in a time of rapidly intensifying climate change. The atmosphere temperature being already 1.2° above the mid-18钍 世纪工业化前的水平，厄尔尼诺现象的峰值温度很可能暂时将大气的热量提高到1.5°。碰巧的是，这个水平只不过是气候科学家确定的气候安全上限的峰值极限（Kate Abnett, "厄尔尼诺现象重现，2023年世界可能面临创纪录的气温", 路透社，2023年4月20日和Ajit Niranjan，"气候变化如何影响厄尔尼诺和拉尼娜周期？", 邓小平, 01/27/2023).
中国进口的88%大豆被用于动物饲料，特别是猪饲料，中国的猪群占世界猪群的一半。猪肉行业满足了中国城市对肉类日益增长的国内需求（Jean-Michel Valantin，" "）。中国、非洲猪瘟疫情和地缘政治", The Red Team Analysis Society，2019年10月14日）。
As China is 1.4 billion strong, it needs to access immense and worldwide resources pools a matter of food security. So, for the Middle Kingdom, importing Brazilian agriculture products is a matter of social cohesion. Reciprocally, from the Brazilian perspective, deepening ties with China is, among other matters, a way to guarantee to Brazilian farmers the imports of the fertilizers they need for the Brazilian agricultural development (Donnellon-May and Porto, ibid).
One must remember that the war in Ukraine triggers a massive disruptions of the European gas adduction from Russia. Thus, Europe needs to access to new resources bases on an already tight market. The problem is that, because of the growing Asian gas consumption, gas prices remain at a high levels.
我们知道，从历史上看，粮食危机有可能成为已经不稳定的国家或大陆地区社会和政治动荡的主要驱动力，人们可能会想，2023-2024年是否会成为重大动荡和政权更迭的年份。正如杰弗里-帕克在其巨著中所说的那样 十七世纪的全球危机、战争、气候变化和大灾难, (Yale, University Press, 2013) notes, during the “little ice age” of the 17钍 世纪，许多政权都是严冬和寒冬所诱发的饥荒、对灾难的不良管理以及对那些在无力养活和照顾人民的情况下榨取税收的政府的叛乱倍增等因素的受害者。
在2022-2023年的深冬，人工智能（AI）再次成为热门话题。这一次，它不是一个人工智能代理的惊人能力、 DeepMind的Alphago (now Alphabet/Google), demonstrated to win alone against a go Master, but the ability to discuss and do “many things as a human” displayed by the latest version of OpenAI (Microsoft) AI model, GPT-4.
As each time disruptive technology spreads, we may complain, struggle against it, or fear it. Yet, it is highly probable (over 80%, although not certain) that AI and more specifically the “GPT-type” of AI, will generate significant changes.*
The first article of the series first presents what are GPT models, ChatGPT and why they matter in terms of occupation and jobs. Then, we test ChatGPT on a specific question: “the future of the war in Ukraine over the next twelve months”.
OpenAI方面也进行了 研究 on the impact on the labor market of its AI models and more generally of Large Language Models (LLMs) (Tyna Eloundou, Sam Manning, Pamela Mishkin, Daniel Rock, “GPTs是GPTs：对大型语言模型的劳动力市场影响潜力的早期观察", ArXiv，2023年3月17日（v1), last revised 23 Mar 2023 – v4). Note, however, that OpenAI also has a corporate and legal interest in promoting LLMs and more particularly its GPTs-models as “general purpose technology.” For OpenAI, to see its GPTs-models perceived as “general purpose technology” would, for example, imply a very strong advantage in the protection of its trademark and in preventing others to use the name GPT (e.g. Connie Loizos, “‘如果OpenAI得逞，"GPT "可能很快会成为商标。", 技术新闻网(TechCrunch)，2023年4月25日）。
“… around 80% of the U.S. workforce could have at least 10% of their work tasks affected by the introduction of LLMs, while approximately 19% of workers may see at least 50% of their tasks impacted…
Our analysis suggests that, with access to an LLM, about 15% of all worker tasks in the US could be completed significantly faster at the same level of quality. When incorporating software and tooling built on top of LLMs, this share increases to between 47 and 56% of all tasks.”
Generative AI is meant to particularly impact “software, healthcare and financial services industries”, and, more largely, the media and “entertainment, education, medicine and IT industry” sectors (Ibid., Goldman Sachs Insights, “稳定的AI首席执行官说，AI将被证明比大流行病更具破坏性“, 31 March 2023).
Second, practitioners of strategic foresight, early warning, risk management etc. are “knowledge workers” and “scientists.” We are thus on the front line of those who will be hit by Generative AI, if Goldman Sachs is right. Hence it is better to make sure AI serves us rather than kill us.
GPT是一个语言模型（自然语言由 兴业银行 (Microsoft) that uses deep learning to generate human-like responses to prompts. The “GPT” stands for “Generative Pre-trained Transformer,” which means that the AI has been trained on vast amounts of text data to understand the nuances not only of the English language but also of other languages. GPT-models are part of the 生成式人工智能 家庭。
Generative AI belongs to the Machine Learning > Deep Learning > Unsupervised, Supervised and Reinforcement Learnings category of AI (for more details on the typologies of AI, see Hélène Lavoix, “人工智能何时为地缘政治提供动力 - 呈现AI", "人工智能和深度学习--正在形成的新人工智能世界", "在现实中插入人工智能", The Red Team Analysis Society).
GPT latest models were trained through supervised and reinforcement learning with human feedback (see OpenAI, “对准语言模型，遵循指令").
Generative AI is thus typically based on deep learning models, which are trained on large datasets of examples in order to learn patterns and generate new content. We have, for example, generative adversarial networks (GANs – Ian Goodfellow et al. “生成式对抗网络“, 2014) and variational autoencoders (VAEs – defined in 2013 by Kingma等人。 和 Rezende等人。 – for an explanation J. Altosaar, “Tutorial – What is a variational autoencoder?“). For instance, a GAN might be trained on a dataset of images, and then used to generate new images that are similar in style and content to the original dataset, but are not exact copies.
The world into which an application such as ChatGPT actually evolves, as well as most GPT-based applications is digital. In other words, most GPT-based apps have no capability to act directly concretely in reality… except if human beings become their actuators (see Helene Lavoix, “人工智能的执行者（1）：在现实中插入人工智能", The Red Team Analysis Society，2019年1月14日）。
Thus, those activities that will be most impacted by GPT-based applications and their likes are those located within the digital world. Actually, we may rephrase this as: the more insubstantial – in the meaning of not having physical existence – our activity, the more likely it will be impacted by Generative AI applications such as GPT-based apps.
The logic is similar to what we identified and explained previously regarding the importance of sensors and actuators for the development of AI (see Lavoix, “人工智能的执行者（1）：在现实中插入人工智能“, Ibid, 2019). This corresponds to what Eloundou, et al. (“GPT就是GPT2023年）发现：
…information processing industries (4-digit NAICS) exhibit high exposure, while manufacturing, agriculture, and mining demonstrate lower exposure [to LLMs and GPTs].
Indeed, OpenAI’s research paper presents a table of impacted occupations, as reproduced here.
Logically, it appears that there are also variations in the degree of exposure of “ideational” activities according to the types of skills most necessary for an activity. Science and critical thinking are the least exposed and impacted by LLMs, whilst programming and writing are part of the most exposed activities:
“Our findings indicate that the importance of science and critical thinking skills are strongly negatively associated with exposure, suggesting that occupations requiring these skills are less likely to be impacted by current LLMs. Conversely, programming and writing skills show a strong positive association with exposure, implying that occupations involving these skills are more susceptible to being influenced by LLMs.”
对于那些会感到失望的人来说，请注意，OpenAI已经向从事以下工作的开发者开放了其模型 用于ChatGPT的插件 that would, for example, allow to “retrieve real-time information”. It is thus a question of time before GPT-based models and applications can be up-to-date. Indeed, the “新宾“, Microsoft new search engine, uses GPT-4 and, 事实上在这里，我们可以访问世界网络。
Even though developments are very fast in the field, for now, our test will be done with what is available: a Generative AI which “knowledge” was cut off in September 2021.
Knowing that most serious methodologies of strategic foresight and warning, indeed any analysis, are grounded in explicit modeling, we shall examine first the capability of ChatGPT to help us with the creation and development of our model (see Hélène Lavoix, “动态风险和不确定因素的建模 (1) : 绘制风险和不确定因素图", The Red Team Analysis Society, 2018 ; Joshua Epstein, Joshua M., “为什么要做模型？", 人工社会和社会模拟杂志》（Journal of Artificial Societies and Social Simulation）。 11(4)12, 2008; 千年项目。未来研究方法,3.0版，2009年）。
However, it’s important to note that the American national interest is just one of many variables that could be relevant to understanding the future of the war in Ukraine over the next 12 months, and it should be considered in conjunction with other factors, such as the interests and actions of other key actors in the conflict.
We further tested ChatGPT by asking if a very specific variable, in that case “the American National Interest” was a variable for our question.
In social science terms, “a variable is a logical set of attributes” (art “变量和属性(研究)” Wikipedia – using Earl R. Babbie, 社会研究的实践, 12th edition, Wadsworth Publishing, 2009, p. 14-18).
The substance of the answer, yet, let us expect that if we correct our question and, eventually, train ChatGPT adequately, then we could – hopefully – obtain the right set of variables with their clear attributes.
It’s important to note that these are just some possible causal links between the variables, and the actual causal relationships between the variables may be more complex and interdependent than these simple relationships suggest.
It could also identify some relationships among variables, but by no means all. To highlight just one missing relationship among many, “conflict escalation” and “escalation risk” are not linked by ChatGPT.
In the above matrix, the row variable represents the cause or independent variable, and the column variable represents the effect or dependent variable. An “X” in a cell indicates a potential causal link between the corresponding row and column variables, and an empty cell indicates no significant causal link.
Again, it’s important to note that these are just some possible causal links between the variables, and the actual causal relationships between the variables may be more complex and interdependent than these simple relationships suggest.
However, with training of the GPT model and effort, it may be possible to obtain something better… to be tested in the future.
As it is, ChatGPT and gpt-3.5-turbo model cannot directly output a good enough model for a strategic foresight and early warning’s question on an issue pertaining to national and international security.
As it is, we estimate that ChatGPT can be nonetheless very helpful to beginners and non-specialists. Iterations and asking increasingly more precise questions appear as the best systematic way to use ChatGPT for variables and causal relationships. We may also imagine the AI can help specialists if the questions asked are precisely formulated – and thus well understood by ChatGPT.
作为一个整体，我们估计这里有足够的潜力来创建一个相应的人工智能助理、 卡尔文, that our members and readers will be able to use. We intend to continue improving this AI assistant. Notably, in the future, further training of the AI for our specific purpose – called 微调 – could yield very interesting results. However, it seems best to wait until the full release of GPT-4 to start such a process, after carrying out a full test again with GPT-4.
*… unless other even more upsetting events take place. The specifics of the changes that will emerge from the use of AI in general, ChatGPT-like AI in particular, may also differ from our current expectations. These two points will be addressed in future articles. Here we concentrate on the highly probable and the evolution generated by ChatGPT-like AI.
**… this task is also used in workshops to open-up the minds of participants, to generate new ideas. For analysts, the very fact to make the cognitive effort to find out variables and their causal relationships enhance understanding and comprehension of an issue. Thus, being able to carry out this task alone or in a group has great value in itself. To see it being completely done by an AI-agent would is thus likely to have serious negative consequences. It will thus be key to creatively integrate the AI tool, once operational, in a way that minimise drawbacks.
碰巧的是，从2020年开始，塞尔维亚一直在经历一个非常困难的时期，面对Covid-19大流行病。在这些严峻的时刻，北京通过提供口罩、卫生和基因组测序设备帮助塞尔维亚（Jean-Michel Valantin，"中国，疫苗和安全的 "健康丝绸之路"", The Red Team Analysis Society2021年2月15日，以及Hamdi Firat Buyuk、Danijel Kovacevic、Edit Inotal和Milica Stojanovic，"土耳其、塞尔维亚、波斯尼亚、匈牙利对俄罗斯和中国的疫苗表示信任", 巴尔干半岛的洞察力，2021年1月22日）。
From the Serbian strategic point of view, this delivery enlarges the military cooperation with China. This cooperation started in 2020 with the delivery of 9 “Pterodactyl 1″combat drones (Jovan Knezevic, ibid).
This Serbia-China military and cultural cooperation deepens with the development of the cooperation between the Chinese tech and AI giant Huawei and the Serbian Ministry of Interior (MOI). Since 2019, the MOI experiments with the installation of a network of surveillance captors in Belgrade, in order to further security. Furthermore, this cooperation also takes place in a couple of “Kosovar communities that are partly outside Pristina’s control”, which implies strong political tensions (Mila Djurdjevic, Sandra Cvetkovic, Andy Heil, “塞尔维亚通过后门在科索沃嵌入中国的窥探工具", 自由欧洲电台、 2022年1月8日）。
These technologies are already widely deployed in Chinese cities, and are used, among others, to establish the “social credit” system. (Jens Kastner, “中国瞄准欧洲，向塞尔维亚出售无人机和华为系统", 日经亚洲，2019年10月2日，和Jean-Michel Valantin，"新丝绸之路上的人工智能", The Red Team Analysis Society, 2017年12月4日）。
This trend became even stronger during the high tide of the Covid-19 pandemic. In these very difficult times, in 2021, Serbia, as many other countries, received sanitary masks, medical material and Sinovac doses from China. It was thus part of the Chinese “Health Silk Road” (Jean-Michel Valantin, “中国，健康丝绸之路和安全", The Red Team Analysis Society，2021年2月15日）。
In the same time, Serbia is having its road and railways infrastructures renovated and expanded by Chinese companies. It also becomes a sub contractant for China since the Chinese take over of the unsafe and very polluting Smedrevo steel mill, which exports its production to China. Thus, it appears that Serbia is becoming a main hub for the Belt & Road initiative in Southern Europe (“中国拥有的钢铁厂使塞尔维亚的城镇笼罩在红尘中，癌症蔓延", 路透社，2021年11月9日）。
That dimension is grounded in an understanding of the spatial dimension of China, in the geographic sense. Space is not only conceived as a support to spread Chinese influence and power to the “outside”. It also allows the Middle Kingdom to “aspirate” what it needs from the “outside” to the “inside”. (Quynh Delaunay, 现代中国的兴起，全球化中的环境帝国, 2014).
这个非常古老的游戏强调的不是控制，而是掌握对手的空间（Arthur Waldron，"中国的军事经典”, Joint Forces Quarterly, Spring 1994). The strategy is to “convert” that space into one’s own. To do so, one has to “surround and conquer” the pieces, i.e. the space of the adversary.
It is true, for example, of Mao’s “revolutionary warfare” against Japan and the nationalist military (Scott Boorman, ibid). As we have seen in The Red Team Analysis Society, it also drives the mammoth “Belt & Road initiative” (Jean-Michel Valantin, "中国与一带一路倡议 "部分, The Red Team Analysis Society).
在这种情况下，中国在2030年大规模投资以成为人工智能领域的世界领导者的经验，肯定会成为沙特党的战略资产。中国在安装人工智能 "城市大脑 "网络方面的经验也是如此（Jean-Michel Valantin, "中国的人工智能革命", The Red Team Analysis Society，2017年11月13日）。
碰巧的是，中国和沙特在人工智能方面的众多合作似乎是沙特2030年大战略和中国 "一带一路 "倡议（BRI）之间融合的一个新层面（Jean-Michel Valantin，"沙特与中国 "一带一路"：大融合", The Red Team Analysis Society，2019年3月11日）。
自2013年以来部署的 "金砖 "战略，旨在确保中国的能源资源、商品和产品的持续流动。这些流动对于这个拥有14亿人口的 "中等国家 "目前的工业和资本主义发展是必要的。(Jean-Michel Valantin, "中国和新丝绸之路--从油井到月球......以及更远的地方", The Red Team Analysis Society报道，2015年7月6日）。从那时起，它吸引了众多亚洲、非洲、中东欧和南美国家的兴趣和承诺。
同时，他们不能按比例提高面包的价格。事实上，购买面包和相关产品的个人和家庭也必须面对通货膨胀和能源价格的上涨。因此，他们不会以高价购买面包，例如法国的3欧元一条长棍面包（以下视频。 法国信息2在荷兰，"Un boulanger de l'Oise ne peut plus payer ses factures"，1月3日）或5欧元一个普通的面包(路透社，"荷兰家族式面包店的面包销售无法支付能源费用"，同上）。面包将成为一种奢侈品，销售量的降低将抵消价格的上涨。
如果大多数个人和企业的行为与我们的第一种情况一样，在国家层面的集体结果可能是可怕的，正如Medef所警告的那样，直接导致欧洲的非工业化，并如我们所看到的那样，导致内腐。请注意，就世界秩序而言，美国可能会发现自己的盟友被大大削弱，其结果可能是有利于加强美国想要征服的秩序（见Helene Lavoix, "美国的国家利益", The Red Team Analysis Society，2022年6月22日）。
另一方面，我们有一家公司注意到全球的情况，注意到长期的紧急情况，即气候变化和地球界限的逾越（见James Howard Kunstler。 漫长的紧急情况谈到 "中国人 "时，他说："中国人 "是 "中国人 "的意思。人类世时代和经济（不）安全 - (1)", The Red Team Analysis Society，2016年；Helene Lavoix，"气候变化、地球边界和国际关系中的利害关系", The Red Team Analysis Society, 2022).
(1) 对于这最后一点："统治者的第三项义务是以有助于......臣民的物质安全的方式行事。......安全，以抵御超自然的、自然的和人为的对食物供应和习惯性日常生活的其他物质支持的威胁。"巴林顿-摩尔，I不公正。服从和反抗的社会基础。 (London: Macmillan, 1978: 21-22)；关于统治者、其义务、社会契约等主题的更多内容，还可参见，特别是马克斯-韦伯。 知识分子与政治在这篇文章中，我们可以看到，"Wissenschaft als Beruf "和 "Politik als Beruf "的原文是1919年；John S. Migdal，S.强大的社会和弱小的国家：第三世界的国家-社会关系和国家能力 (Princeton: Princeton University Press, 1988); John Nettl, "The state as a conceptual variable," (国家作为一个概念变量). 世界政治谈到这个问题的时候，他说："我想说的是，我们都有一个共同的目标，那就是把我们的工作做得更好。 利维坦的诞生。中世纪和现代早期欧洲的国家和制度建设.剑桥，英国；纽约。剑桥大学出版社，1997；Helene Lavoix，"Identifier L'État Fragile Avant L'Heure:预测指示器的作用"，编辑卷。 脆弱的国家和社会 (法国发展署和法国外交部） - 2007年1月。