The European Commission has issued guidelines to increase the ethics of artificial intelligence

accepted the work of a group of experts established to increase trust in artificial intelligence and published new guidelines on the subject ● Dr. Aya Sofer, IBM: "As artificial intelligence is integrated into more and more technological platforms, the need to increase trust in it increases"

The European Union Commission recently presented steps it will take to build trust in the use of artificial intelligence, relying on the work of a committee of experts.

Following the recommendations of the group of experts appointed in June 2018, the Commission launched a pilot that will begin in the summer of 2019 to ensure that the ethical guidelines for the development of artificial intelligence can indeed be implemented in practice.

The Commission invited industry people, research institutes and public authorities to examine the detailed assessment list of the expert group that supplements the guidelines.

The plan presented last week is part of the Artificial Intelligence Strategy Plan from April 2018, which aims to increase public and private investments in the field of artificial intelligence by at least 20 billion euros every year in the next decade.

Help detect fraudulent threats and cyber incidents
Artificial intelligence could help a large number of sectors such as health, energy, vehicle safety, agriculture, combating climate change and financial risk management.

Artificial intelligence can also help detect fraud threats and cyber incidents, allowing law enforcement agencies to fight crime more effectively. However, this ability also brings new challenges to the future of employment, and raises legal and ethical questions.

The Commission realized that this would require three steps - defining the key requirements for reliable artificial intelligence, launching a large-scale pilot and receiving feedback from stakeholders, and the third step - building an international consensus for human-directed artificial intelligence.

Seven elements for achieving reliable artificial intelligence
According to the Commission, reliable artificial intelligence must respect all applicable laws and regulations, as well as a series of requirements listed in the expert committee's report:

• Human feedback and supervision: Artificial intelligence systems should support the basic rights of citizens and not harm them, limit or mislead the humans who use them.
• Robustness and safety: A requirement that the AI ​​algorithms are secure, reliable and strong enough. Intelligence requires that algorithms be secure, reliable and robust enough to deal with errors or inconsistencies during all stages of the life cycle of artificial intelligence systems.
• Privacy and control over the data: citizens should have full control over their data, while the data concerning them will not be used for harm or discrimination.
• Transparency: It must be ensured that it will be possible to guarantee the traceability of the artificial intelligence systems.
• Diversity, non-discrimination and fairness: AI systems should consider the full range of human abilities, skills and requirements, and ensure accessibility.
• Social and environmental welfare: Artificial intelligence systems must be used to change society for the better, and improve sustainability and environmental responsibility.
• Accountability: Mechanisms must be put in place to ensure accountability and responsibility for artificial intelligence, intelligence systems and their results.
  Recently, one of the pioneers of artificial intelligence, Prof. Yehoshua Banjo, a former Israeli who studies the field at the University of Montreal and a recent winner of the Turing Prize, was interviewed for the journal Nature. Banjo is one of the drafters of the Montreal Convention designed to protect the ethics of artificial intelligence

In response to the question of whether he knows companies that use artificial intelligence in an irresponsible way, Prof. Banjo said that "there are many such companies and there may be many more, so flags should be raised before the phenomenon spreads. Most of the worrisome things do not happen in the light of day but in military laboratories, security organizations , and in private companies that provide services to governments and police.

As an example, Prof. Banjo gave the killer drones, which raise moral and security questions, another example is the surveillance, which can be debated as to whether it also has social benefits - but the danger of exploitation, especially in authoritarian governments, is palpable. In principle, AI is a tool that can be used by those who have power to preserve power and increase it. Another area is that artificial intelligence can increase discrimination and biases such as gender or racial biases because these are found in the data that the technology is trained on that reflect human behavior.

The regulations were drafted with the assistance of IBM. Aya Sofer, global vice president for artificial intelligence technology research at IBM, also referred to the guidelines and said that "as artificial intelligence is integrated into more and more technological platforms, the need to increase trust in it and to ensure that the basis on which it rests is free of bias, fair and transparent. The European Union guidelines document places A global benchmark in the effort to promote artificial intelligence in an ethical and trust-inspiring way - an effort that Shiv will manage Proud to be a partner in it. The document reflects many principles that BAM has been implementing for some time, and an example that other countries in the world can learn from."

to the notice of the European Union

• The ethics of AI: "Leave the human in the loop"
• The Montreal Declaration: Why we must develop artificial intelligence responsibly
• "The fact that something can be done with artificial intelligence does not mean that it should be done"