#machinelearning…

Categories Programowanie

#machinelearning #programista15k #programowanie #bekazlewactwa

Podobno Polska, poprzez zasciankowość i brak postpowego myślenia zablokowała przyjęcie unijnego dokumentu o sztucznej inteligencji …bo w jego tekście użyto słowa „gender.

Zobaczmy zatem:

https://www.consilium.europa.eu/media/46496/st11481-en20.pdf

Dużo górnolotnego pieprzenia o prawach, zagrożeiniach i ludzkim uzywaniu AI (czyli technologii, która jeszcze nie istnieje), a okazuje się, że chodzi o cenzurę. Dokument o AI, ale musieli wrzucić słowa-wytrychy: przemoc wobec kobiet, gejów i lesbijki i mowę nienawiści. Nawet COVID-19 tu wrąbali.

Zwykla lewacka cenzura. Dokument wzywa kraje EU, zeby tworzyly rozwiązania prawne dotyczące stosowania ML/AI z poszanowaniem praw mniejszości, równości płci, Konwencji Instambulskiej (po co ją tu wcisnęli??!!). Mowiąc krótko: rozwój AI tak, ale tylko z kagańcem lewicowej ideologii. Nie dane, nauka, wzorce. ML/AI bedą musiały najpierw wziąć poprawki na mniejszości, bias ze względu na płeć i to, co w danym kraju uchodzi za mowę nienawiści.

Czyli nie przyjęliśmy dokumentu, który de facto blokuje rozwój technologii AI.

2. […] We reaffirm our commitment to the
prohibition of any unlawful discrimination based on any ground as enshrined in Article 21
of the Charter. We also reaffirm our commitment to the accession of the EU to the European
Convention on Human Rights and Fundamental Freedoms as enshrined in Article 6(2) of the
Treaty on European Union. Moreover, the EU will cooperate with the Council of Europe in
the promotion of gender equality and women’s rights and on combating violence against
women and domestic violence: the Council of Europe’s ‘Istanbul Convention’ has been
ratified and is being implemented by 21 EU Member States.

20. It is necessary to fight illegal content online, including hate crime, but at the same time to
protect the right of freedom of expression and the right of information. There is a need to
discuss in which circumstances and to what extent the results delivered by AI systems to
delete specific contentshould require meaningful human intervention. In sensitive areas,
service providers should make their AI systems transparent, for example by granting
research access to them.

21. AI-based applications present opportunities for assessing fundamental rights compliance,
including in the field of equality. Decisions based on algorithmic systems of predefined
rules and accurate and adequate data that is fit for purpose could be less prone to biased
results than human-made decisions. At the same time, the use of AI may in certain cases
also perpetuate and amplify discrimination, including structural inequalities. Equality is of
particular concern in this regard. We take note of the opinion on ‘AI – opportunities and
challenges for gender equality’ adopted by the Advisory Committee on Equal Opportunities
for Women and Men. Safeguards are also needed regarding linguistic diversity and the
accessibility of services to citizens in the official languages of the EU.

22. Data used to train AI systems therefore have to be accurate and adequate for their purpose
and potential biases have to be addressed while allowing for sufficient flexibility in
Research and Development for the further development of these systems. In this respect, we
underline the importance of the principles of equality and non-discrimination in the design,
development, deployment, use and evaluation of AI, particularly in systems integrating
machine learning, and of ensuring that such systems are subject to adequate safeguards and
oversight, including market surveillance.