The spaces for regulatory experimentation in the proposed EU Regulation on artificial intelligence – in particular the articles 53 and 54 – assign to Member States the control function of AI systems born in digital environments, based on web 2.0 technologies.
Within these spaces, in the present of the present, man is in the company of generic algorithms and still manages to exert his influence and define its semantics.
For the near future, it is likely to predict a series of connected IoT devices, estimated at 38.5 billion in 2025, functional for self-learning algorithms and predictive computational engines in qubits regardless of the relationship with the man.
This technical evolution proves it vulnerability on the ex ante evaluation and control of some high-risk AI systems and opens a challenge on future areas to be regulated.
If the rules follow the technological evolution
At the dawn of the Metavers, the EU Regulation on Artificial Intelligence, although the most harmonious of legislative instruments at European level, preserves the principle of territoriality of Member States in the management of regulatory testing areas for the development, testing and validation of innovative AI systems, in the establishment of national certification bodies, in the establishment of notifying authorities called for and notify certification authorities. A government structure whose degree of interoperability and convergence between the same authorities has yet to be tested who runs the risk of being influenced by the level of skills of the resources called to work in them and with a bureaucratic-centric vision in the conditions of operation.
Tailor-made ERP for SMEs: from problems to solutions for the shortest path [infografica]
Teilhard’s quote “at increasingly complex evolutionary stages corresponds to an ever-increasing level of consciousness in beings” opens the fundamental reflection of this exercise of the noosphere: artificial intelligence systems originate in fully digitized environments. In the present, systems operate on data and information released by people, through simple navigation or extended reality experiences, and by objects, through the IoT. They are transnational systems, not yet Metavers — beyond the Universe — but universal. In the present of the future it is very likely that the data population of objects and self-learning algorithms will have the dominant weight in the configuration of digital spaces and in the determination of the technological directions of AI systems.
The deluge of data and technological evolutionism bring equitable and democratic reflections on the relationship between time and latency and between territory and space of regulatory spaces to the attention of “evolutionists”. Artificial intelligence technology ecosystems, including Made in Europe, are faster than regulatory legislative tools. Technological ecosystems originate and exert their relationship mostly within a digital space, where territoriality is moving from a global level of porters to a distributed technological decentralization.
And pprojects funded by the Horizon Program
This reflection is confirmed by the rapid technological evolution of the projects financed with the Horizon Program. Projects such as: AI4EMS, an algorithm capable of supporting physicians and emergency operators in managing the risk of cardiac arrest outside the hospital. Taking advantage of advanced speech analysis, the algorithm allows faster recognition of cardiac arrest outside the hospital, less than 3 minutes compared to the human operator, and more accurate reaching 95% of the “success” of the diagnosis. XAIDA – Artificial intelligence for detection and attribution – for the development of new AI techniques capable of investigating the causes of events and extreme atmospheric dynamics, climate modeling, random paths to the origin of these events, such as now cyclones, heat waves, floods and droughts. The technology used is machine learning from machine learning and is experimented with 3 three main inference paradigms: deep learning, generative models, and Bayesian (probabilistic) nonparametric models to generic algorithms. FANDANGOa solution that provides for the management and analysis of intersectoral Big Data in an interoperable platform, capable of recognizing false news and post-truths, to support traditional media industries to face the new “data” news economy with better transparency for citizens under the prism of responsibility, research and innovation.
In the present of the past Europe has built AI4EU, the first European on-demand artificial intelligence platform that brings together resources, tools, knowledge and algorithms from all member states that drive the needs of the former artificial intelligence community that includes European ELSE aspects (ethical, legal , socioeconomic) in a duopolized global context. for North America and China. The platform reduces technological and ecosystem gaps by creating services that facilitate their use and adoption by end users and researchers; defines sustainable processes and governance structures, business models, and third-party financing for AI projects. Today the platform brings together more than 80 organizations between universities, research centers, companies and associations, has 16 ecosystems or development projects underway, including TAILOR, Human-E-AI-net, AI4Media.
Towards mixed initiatives and hybrid decision-making systems: man and AI become operational partners
In the present of present, technological development is driven by a human empowerment approach in collaboration with algorithms and the research environment. A regular prism Man-Machine-Environment. The Horizon Europe program is proposing two functional and complementary calls for the construction of an AI ecosystem where the relationship between humans, AI systems and Reciprocity guides the process of technological evolution.
AI for human empowerment
AI for human empowerment (TOPIC ID: HORIZON -CL4-2022-HUMAN-02-02) aims to carry out mixed initiatives and hybrid decision-making systems, by building the next level of perception, visualization, interaction and collaboration between men and AI systems. . Man and AI Systems become operational partners, responsible for strengthening the element of trust and transparency of AI systems when defining performance and mixed decision-making processes. The element of trust together with the strengthening of the ethical principles of digital technologies, inclusion and sustainability are the basis of these new forms of collaboration. One of the main activities will be to find new ways to involve citizens, or their representatives, in the development of AI. With these premises, Europe aims to develop a global technology that adheres to European principles and standards, to increase the presence of European actors in the market of emerging technologies.
European Network of Centers of Excellence in AI
European Network of Centers of Excellence in AI (TOPIC ID: HORIZON -CL4-2022-HUMAN-02-02) aims to strengthen the European AI community, monitor the AI4EU platform, improve existing research capabilities and achieve critical mass through closest networks of European centers of excellence in AI. . Proposals should develop mechanisms to connect centers of excellence in AI, mobilize researchers to collaborate on key technical and sectoral or social challenges, increase the impact of funding to move faster in joint efforts rather than working on in isolation, with fragmented and duplicated efforts. The topics around which the new ecosystems of centers of excellence will have to be established are: 1. Next-generation artificial intelligence, which covers fundamental research in artificial intelligence and machine learning; 2. Scientific research and priority technologies that appear in the latest SRIDA (Strategic Agenda for Research, Innovation and Deployment of the PPP AI, Data and Robotics) and that involve the centers of excellence of the network selected with previous Horizon calls ( 2020).
In the transition from web 2.0 to web 3.0Europe, with the joint initiative EURO HPC, commissions part of the exploration of high-impact algorithms for the next European exascale supercomputers to a new ecosystem of researchers from different domains. In the European high-performance computing joint venture (TOPIC ID: HORIZON-EUROPHPC-JU-2022-ALG-02-01), the solutions to be developed are based on an impact assessment on reducing resource consumption HPC that takes into account the current use model and potential use cases of algorithms to explain the concept and design of a fundamentally new and innovative algorithm; define the process of implementation and validation of the new European hardware and exascale supercomputer architectures; demonstrate the expected impact in terms of relevant use cases with a large user base, recovered computation time, or reduced time for the solution. EURO HPC is expected to compete with Google’s Quantum Artificial Intelligence Lab.
Focusing regulation on a family of technologies guarantees respect for and compliance with the values, fundamental rights and principles of the Union with a “territorial border” and limited to the “present of the present”. It’s a lot but more can be done. There is a lot of noosphere and expectations within the paradigm shift that traditional legislative instruments seek to regulate. But “from the succession of evolutionary phenomena” (Theilard) and from the expectation of “the present of the future” (St. Augustine) it is legitimate to ask a new strategy in defining legal frameworks.
The superposition or hybridization between the physical space and the digital space that arrives with the Metavers, the autonomous organizations decentralized by DAO, the plots or the quasi-ordinaryity of the NFT, the volume of data generated by self-learning algorithms predicting digital realities, calls Europe is involved in renewing its role as an exporter of democracy, assuming the political commitment to design a “basic legal framework“That protects the person and their values, the institutions and their functioning, in general the European digital community in the own socioeconomic interaction of the metavers.
With regard to the performance of the implementation of the regulation, in compliance with the principle of subsidiarity, for the notifying authorities of the Member States, it would be interesting to raise the hypothesis of building a dedicated blockchain, each authority a block, for inform art. 46 “Information obligations of notified bodies” in blockchain to disintermediate notification times and spaces. So much so that the highest risk systems are recognized and certified as such immediately by the 27 member states.
@ALL RIGHTS RESERVED