← All statements
Plenary speech

Brando BENIFEI

The whole world is watching us as alarms follow one another about artificial intelligence in general, the risk of extinction of the human race by sentient algorithms, alarms often launched by the same Big Tech that fear what they have created.

Artificial Intelligence Act (debate)official record ↗

Read the full statement ↓

(IT) Mr President, ladies and gentlemen, the whole world is watching us as alarms follow one another about artificial intelligence in general, the risk of extinction of the human race by sentient algorithms, alarms often launched by the same Big Tech that fear what they have created.

In the meantime, Europe has moved forward, without stopping, and in recent months has worked hard to propose a concrete response to the risks that already exist, not the future ones, of powerful technologies such as artificial intelligence. They told us: This will stop innovation, so Europe will never be competitive. While today many of these same people are right: Rules are necessary.

Today, therefore, we are here to ask Parliament for a mandate to start negotiations on the world's first law on this matter. Our position contains many improvements to the Commission proposal, which clarify and update the text, promote legal certainty, avoid duplication with existing laws, stimulate innovation, while keeping the human being and his fundamental rights as a staple.

It must remain at the heart of the European approach to artificial intelligence. I will mention just a few of the key points of our proposal: a clear and balanced definition, in line with the work of international organisations; a chapter to regulate systems such as ChatGPT, previously excluded from the text, to obtain transparency and mitigate the risks of discrimination and the generation of illegal content; in addition to watermarking to make the content produced by generative AI identifiable, which, although already provided for in fact by Article 52, I think will need to be even more clearly specified in the trilogue. But we haven't even forgotten the risks to the environment in training algorithms with billions of data and parameters. We have introduced a requirement for companies and public authorities using high-risk AI to carry out a fundamental rights impact assessment based on the specific context of use, as an additional safeguard for vulnerable groups.

It should also be recalled the addition of the right to a judicial remedy, to an explanation, to be represented by collective action in case of negative impact on health, safety or one's fundamental rights, and the addition of the environment, democracy and the rule of law among the objectives to be protected, in addition to the obligation to discuss the deployment of artificial intelligence in the workplace with trade unions and to inform the workers involved, as well as the expansion of practices considered unacceptable, such as predictive policing, biometric categorisation, indiscriminate collection of facial images from the web, emotional recognition in some contexts and opposition to the use of biometric recognition in public spaces for surveillance purposes.

The use of these cameras has no proven benefit on the safety of citizens, while the risks of abuse, discrimination and inaccuracy are proven, especially on some people, women, non-white people, children. The introduction of these technologies poses risks to the viability of our countries, to their democratic viability, and that is why we will fight with the Council in the coming months to defend this approach and counter dangers to our freedom.

I think of the worrying words of some interior ministers, including the Italian Piantedosi. We want artificial intelligence to unleash all its positive potential for creativity, productivity and the improvement of everyday life. For this reason, we have also engaged in an intense work of confrontation with legislators around the world to find common approaches.

This is why we support the work onAI Pact to anticipate the effects of the Regulation and theAI Code of Conduct for generative artificial intelligence. But let us also make a clear warning: all approach and adaptation activities are useful, as long as they do not intend to replace the binding rules that we are going to establish. We will not be influenced by attempts that could come from large corporations or government institutions and not to dilute and weaken this effort to make our democracies stronger in the context of technological transformation. We will do our best in the interests of Europeans together with the Commission, which I thank, together with the Council, together with all my colleagues.

Text as published by the European Parliament, in the Parliament's own English rendering of the original language.