The strong interest in artificial intelligence (AI) was confirmed in 2021 with massive investments in start-ups and software companies. The United States account for two thirds of global investments and leads in the number of unicorns followed by China, the United Kingdom and Israel. To date, France has two unicorns.

Investors are attracted by companies whose valuations can double in a short period of time. The massive investments also reflect emerging trends among users. According to a recent IPSOS and the World Economic Forum survey, two-thirds of adults believe that their lives will be profoundly changed by products and services using AI within 3 to 5 years, 64% consider that they have a good understanding of AI, 50% consider that the benefits of AI are greater than the disadvantages while 39% remain cautious or even anxious.

Only 13% of respondents believe that rights and fundamental freedoms will be affected by these technologies. This proportion is even more surprising than some AI applications directly impacting individual rights and fundamental freedoms have been generalized and tend to be almost normalized in a large number of countries, such as facial recognition in public spaces or image detection technologies supposed to identify illegal behaviour (e.g., Apple's CSAM).

In that context, 2022 trends will echo and reinforce the 2021 trends with: 

• Hyper-automation of business and IT processes to streamline the value chain, automatic decision-making systems, (hyper)-personalization of customer experiences, rapid advances in natural language processing (NLP) reinforced by the development of multimodal AI, progresses made with semiconductors reducing the environmental cost of AI, anonymization and confidential treatment of data, preference for public and state’s actors for open source and opt-in systems where individuals hold the sovereignty over their data, which can be made accessible, or not, and for which they can be remunerated by crypto-currency. Additional unforeseen technological breakthroughs will occur as well. 
• Development and deployment of sectoral AI solutions especially environmental AI applications with the influence of the climate emergency agenda, the European Green Deal and international conventions.
• Implementation of AI legal and regulatory frameworks that will complete the existing or currently adopted legislations ("Digital Markets Act" and "Digital Services Act") in terms of data. In 2021, Europe continued its normative efforts to guarantee the fundamental values and rights of the European Union, legal certainty and develop safe, legal and trustworthy applications while facilitating innovation with the proposed Artificial Intelligence Act. China has established a new regulatory framework that enables the country to maintain its technological lead, confirm its extremely strong market presence, and build a legal framework in line with Western concerns and ethical debates. The Chinese law on the protection of personal information ("LPPI") would quickly be complemented by a law proposal on AI. The current draft suggests the establishment of a control over algorithms. In that sense, the tests that will be conducted by Chinese companies to comply with the algorithmic explainability requirement will be interesting to watch. All those initiatives create a ripple effect and an international dynamic that could lead to 2022 being the long awaiting year of the balance of power between states and Big Tech. 
• States will have to make increasingly decisive, even critical, strategic choices to continue their modernization without losing their state power to private actors or harming the social fabric threatened by the growing weight of metaverses, new digital forms of political destabilization and the increased disconnection of citizens from their institutions due to the growing weight of intermediaries exercising almost public service missions without abiding by the constraints or obligations related to such missions. AI can serve as a catalyst to better understand the current challenges if it supported by holistic strategies grounded in the general interest, climate issues and fundamental rights and if deployed through sectoral tactics allowing contextualized cost-benefit-risk analysis. Returning to a distinction between safety and security would be beneficial and necessary. Lastly, the data rationality and objectivity rhetoric calls for more impact assessments
• Decisions of state actors will be influenced by the supremacy and geographical concentration of technological know-how. International cooperation and international diplomacy will face growing challenges related to data sovereignty, the right to be forgotten, the consent of vulnerable populations and the delicate balance between international aid and digital submission. 
• Ethical AI issues (discrimination, bias, decolonisation of AI, data quality, transparency, fairness and algorithmic justice, etc.) will continue to fuel iterative processes of good corporate governance. 

One question that needs to be addressed more systematically is whether we really need more technology and information. Recently, Oxford University organized a debate with a robot debating with itself about the usefulness of AI. His conclusion was that we are not smart (or evolved) enough to make the right moral or ethical choices, and therefore we should avoid the race for AI by not using AI or by having an AI system implanted into our brains.

Photo by Kelly Sikkema on Unsplash