Shaping the future with artificial intelligence
Learning, perceiving, understanding, reasoning, making decisions, anticipating… these are all faculties that only the human brain seems capable of. And yet, they are all claimed to be possible with computer systems that fall within the concept of artificial intelligence (AI). This generic term refers to a broad and varied range of systems of varying levels of complexity. One thing is certain: AI’s applications are revolutionising almost all swaths of our societies and our economies.
Artificial intelligence is one broad trend that no one can ignore. Going beyond the headlines, AI is a booming market – its value is expected to have risen from 86.9 billion dollars in 2022 to 407 billion in 20271. AI is big business but also a technological revolution whose repercussions are likely to equal or exceed those that were felt at the advent of the Internet, back in the day. Indeed, few sectors will be unaffected by one form or another of AI. Why is that? Perhaps quite simply because AI meets the needs of our societies and our economies.
Machine learning and deep learning: when machines learn how to learn
Take the healthcare sector. AI has been focused on it for several years now, thanks to its ability to integrate, manage, contextualise, and analyse vast volumes of health-related information or data from the medical literature, for example. As far back as the early 2010s, IBM Watson joined the Memorial Sloan Kettering Cancer Center in New York to help diagnose cancer patients2. As the technology was not yet there, the experiment was inconclusive. But progress since then has been breath-taking. The Curie Institute, for example, has developed an AI algorithm able to identify cancerous tumours of unknown origin3. And AI is gradually carving out a place with healthcare professionals for tasks such as analysing medical files and assisting in diagnoses or clinical research. The development of AI machine learning, deep learning, and image recognition will accelerate in these fields in coming years. Likewise, the use of AI will facilitate the development of both new therapeutic drugs by the pharma industry and personalised medicine and patient follow-up.
AI’s machine learning and deep learning capabilities – and its ability to anticipate based on data and scenarios – may also find broader applications in the financial sector. Analysing market risks and trends, identifying potentially fraudulent behaviours or transactions, and so on – the range of applications is vast, but not without risk, as the OECD recently noted: “The use of the same models or datasets can lead to convergence and herding behaviour, increasing volatility and amplifying liquidity shortages in times of market stress.”4
Chatbots and generative AI are rethinking how we communicate and create
ChatGPT has been a roaring success. By February 2023, just two months after it was released to the general public, it already had more than 100 million active users. This highlights the attraction of another AI branch: chatbots, particularly those that integrate machine learning capacities, as is the case of ChatGPT. Chatbots have already been a game-changer in many sectors. They are in place on many websites to guide and support users and answer their questions. One example of this is Hopla, a chatbox being prepared by Carrefour, a French retailer, that is based on OpenAI technology5. It would be used to help shoppers on its French platform choose items based on their budget and food constraints, for a given number of persons, while providing recipe ideas. Still at the prototype stage, Hopla could transform retailing. “This is undeniably the start of a revolution and the transformation of our businesses and our organisations, and the way that we speak to our customers”, saidAlexandre Bompard, Groupe Carrefour’s chairman and CEO, in an interview with BFM Business, a French radio programme. This use of technology could revolutionise the customer path as the AI involved is finetuned and made more adaptable.
These faculties make it a promising tool in education, as well. “AI adaptive learning systems can quickly identify when a student is struggling and then provide more or different support to help them succeed. As the student shows that they have mastered the content or skill, the AI tool provides more difficult tasks and materials to further challenge the learner”, the World Economic Forum noted in a paper on AI use cases for rethinking learning6.
Meanwhile, a recent US study found that 30% of students surveyed had already used ChatGPT to write research papers7. This points to generative AI’s potential creativity or its capability of producing new data via machine learning. Generative AI programmes, such as Midjourney or DALL-E, can be used to create music, text, images, and so on. This pushes back the horizon on creativity but also raises many issues, such as compliance with intellectual property rights or the question of whether generated content is real and accurate8.
Among AI’s many applications, AI-augmented automation is improving its performances and broadening its reach. This is the case in manufacturing, which has long been familiar with automation, robotisation, and industrial vision. It is now accelerating its transformation through broader and broader use of AI, making robots that are not only better-performing and adaptable but that can also truly work alongside humans. This would significantly lighten the burden of certain arduous or repetitive tasks, while enhancing performance and agility.
Mobility is one of AI’s best known and furthest advanced applications. Fully self-driving cars are not yet part of daily life, but most modes of transport do integrate intelligent applications to secure and facilitate mobility or to optimise energy use.
Supervising and regulating
That being said, the development of various forms of AI is raising many questions and even some opposition. Back to the example of ChatGPT: in February 2023, Italy temporarily banned chatbots until it could ensure that users’ personal data was being used in a manner that complied with the General Data Protection Regulation (GDPR), a ban that it lifted in April9. Beyond the matter of compiling data and putting it to use, which is an issue common to all AI applications, AI poses a myriad of ethical questions (regarding transhumanism, for example), along with social and legal ones. To wit: who (or what) is criminally liable if an AI system kills or injures a human being, accidentally or on purpose?
This question is far from academic, given that defence is at the cutting edge of AI. There are indeed many military applications, including simulation, smart drones, analysis systems, cyber defence, or even battle-ready robots, and they are generating heavy investments. During budget negotiations last February, the Pentagon asked for 1.8 billion dollars to develop AI within the US Army10. And in June 2022, NATO announced a USD 1 billion investment fund to support startups that develop “priority” technologies, such as artificial intelligence, data and automation11.
Meanwhile, calls and proposals to regulate the use of AI and autonomous systems are being heard in all sectors, and especially defence. In February, the US State Department put forth a set of principles meant to govern responsible use of AI: “Military use of AI capabilities needs to be accountable, including through such use during military operations within a responsible human chain of command and control. A principled approach to the military use of AI should include careful consideration of risks and benefits, and it should also minimize unintended bias and accidents.”12
Transforming and guiding
Regardless of the challenges facing AI, there is universal agreement that all sectors and activities will be affected by its deployment, starting with the workplace. Will AI create more jobs than it destroys? Various figures are being waved about, but the only consensus that has been reached is that some professions will go by the wayside or forced to evolve, while others will be born or developed further. However, when used properly, AI will make some jobs less arduous, and will replace humans in some sectors, such as personal care services, which have a hard time hiring. It will give rise to new skills and new professions.
As with the Internet, one response to these issues is investment – not only to develop and implement AI technologies, but also to provide support and training and make them as broadly accessible as possible. The European Union plans to invest a total of 20 billion euros per year in AI13. And, as noted in an EU report from 2020, the top source of investment by European countries in 2020 (34%) was in developing skills and transforming work through AI14. Meanwhile, there are burgeoning private-sector initiatives and investments. After investing 1 milliard dollars in OpenAI in 2019, Microsoft raised its ante 10-fold early this year, with a new, 13 billion investment. This is no mere financial investment, as Microsoft is also offering its services to the company that created ChatGPT, including its calculation power and programming interfaces15.
While AI will gradually transform the way we work and our societies in general, it will also help address the crucial challenges of our time. In 2022, the UN launched the “AI Initiative for the Planet”, to encourage international cooperation through conferences and public-private partnerships between researchers, scientists, policy-makers, entrepreneurs, startups and private-sector actors16. Their goal: to show how AI can help address the challenges posed by climate change by assisting transport, agriculture and industry in calculating and reducing their greenhouse gas emissions, by enhancing energy efficiency in buildings, by optimising renewable energy production, and by enhancing the management and prevention of climate-change-related risks. AI is a key tool in all these fields, and will be one way to make the world more responsible and sustainable.