In our transforming world, digital technology has the critical mass to push our frontiers and release unlimited potential. As the wave of digital transformation soars high, improving our lives, industries and economies, we must not overlook the risks that technologies and innovations pose on the fairness and cohesion of our societies, and our rights as European citizens.
Through the construction of vast digital infrastructures, which track, monitor and collect personal data at scale, major online platforms are commodifying our data. An entire ecosystem of apps and companies controls huge data flows and ‘treats’ technology users as data suppliers via the imposition of default choice architectures. The current form of ‘computational governance’ in which we are ruled by algorithms and predictive analytics which make use and trade of individual human experiences to produce predictions of human behaviour has largely remained unregulated. We can still be in control of safe AI, trustworthy and complimentary to humans.
AI is at the core of this wave of digitisation; the power it carries must be tamed before it is unleashed in its next form of super intelligence. Manifested in both software and hardware with intelligence expressed in lines of code and powerful automation, it is already being used to solve complex problems, discover patterns or predict traffic, and make suggestions to users. AI is also tested for completely autonomous vehicles, as deep machine learning could make decisions faster, safer and better than people. Progress remains incremental, however the more we advance our understanding and the further we develop this technology, the closer we reach the point where AI leaps from narrow to general, and then super intelligence. Narrow AI is already making simple suggestions and decisions at the service of people; however, the next level of AI sophistication can completely omit humans from the process and reach levels that humans will not be capable to understand or control anymore.
As most of human activity and industry has moved online during the pandemic, and we become more dependent on an increasing amount of digital services, a series of critical questions arise: How can we build data infrastructures that recognise the role of data as a public good? Can we foresee an impartial system of data governance that could reconcile the astonishing potential of these technologies with their significant human downsides? Do we as citizens know that when we use a search engine to find something online, the engine is learning by searching us as well? Does the collection and processing of behavioural data by these technology platforms threaten to dismantle democracy and undermine the rule of law?
To tackle these questions we need to come up with novel, innovative responses, so that we make our future better than the grim outlook predicted in Black Mirror, the popular Netflix series. We have to be aware and cautious in designing the appropriate framework for AI; intelligent systems are used on the pandemic front for rapid COVID diagnosis and prognosis based on AI-enabled CT scanning analysis. AI systems are used now to upgrade our mapping systems, the management of resources, to make our mobility smarter, our agriculture more precise, with predictive maintenance and smart sensors that can contribute to a greener future, saving time, energy, and resources.
As the impact of exponential technologies grows, a new framework is needed in Europe to harness its benefits and mitigate its risks. This framework must be built on ethical principles and binding standards, which elevate people’s trust in AI and ensure that in the digital age, people co-exist with intelligent systems without fearing exclusion, manipulation, oppression or discrimination. Retaining freedom of choice in a human-centric AI that would prevent brain computer interfaces challenging the nature and future of humanity. In contrast to the trends of the Fourth Industrial Revolution towards inequalities and dehumanization, technology and innovation best practices need now to be bent back towards the service of humanity, and Europe could lead as a global rules and standards setter for the Fifth Industrial Revolution.
The European principle-based framework for AI systems must translate and establish by law, respect to our rights in the digital age. Progress has been made with landmark initiatives such as the GDPR, however more work is needed in this space to establish legally binding rules which provide developers and innovators with more legal clarity, and award individuals stronger protections of their rights and freedoms. At the foundational level, this framework must guarantee higher transparency and accountability, define the liability of AI systems, establishing standards to trace, audit, explain, appeal and reverse decisions made by AI during its entire lifecycle.
The rapid development of automation in Europe must not reflect mistakes of the past; AI algorithms and systems must be trained on diversified sets, and their objectives have to be clearly defined and controlled to avoid risks of bias and discrimination or data poisoning. Decisions made by AI must be aligned with the collective ethical fabric that defines Europe throughout the lifecycle of intelligent systems and follow clear red lines, such as the risk-based approach that would, for example, completely ban research in autonomous lethal weapons or conscious AI. The high-risk applications should alert us of the fragility of our common values, as well as our rights as citizens, our responsibilities as policy-makers and our obligations to future generations.
Strong protections can enable the development of ethical AI in Europe, and force gatekeepers to respect with reciprocity the new architecture of data governance. The data ecosystem is ground-zero for AI innovation in Europe. Data is the source of learning for AI systems and therefore very important for AI to become smarter. Privacy by design, consumer protection, and product safety will be at the top of our agenda since we will not compromise privacy for safety. More options available to users, clear consent, and rewards for the use of our data, with stronger security standards, as in cases of extreme personalization and micro targeting, would enable fairer AI in this evolving ecosystem, and inspire trust for European citizens. The EU has expertise in leading efforts to establish global norms and standards for data processing, which it can leverage in collaboration with international partners to achieve global consensus. In the European Parliament, we are currently working on an ambitious agenda to enable better, rules-based data collection and safer mechanisms to process, store, exchange, and transfer data. Standardisation is necessary to ensure the high quality and safety of data, alongside the establishment of data spaces governed by rules addressing the divergent levels of inherent risk of different sectors and technologies. Coupling these efforts with a secure digital ID can be instrumental in rethinking the current data ownership models.
In the EU, a significant portion of the budget and other funding instruments is earmarked for the development of digital skills, and digital education systems fit for purpose, as we must prepare the grounds for a seamless transition of our societies and economies into an era powered by intelligent systems. At both the national and the EU levels, progress is necessary to calibrate education and professional training with the needs for digital skills, to enable greater understanding of AI across generations, and stimulate career paths into future jobs of AI innovation and research.
Our vision is to foster technology that creates value for all, and enables our society to progress without impeding on the rights of Europeans. Striking this sensitive balance is a complicated endeavour, however the acceleration of AI development, and the reach of its impact, implore us to get it right from the outset. The EU is a global champion of privacy and data protection in the digital age. This achievement must be leveraged in establishing a safe ecosystem, where trustworthy AI enables us to achieve our objectives with regulatory control, and establish a fair and equal society fit for the future.