Artificial intelligence is a hot topic – technologies are polarising, media attention is increasing. In our digital press review, we compile important articles on the subject monthly in order to provide an introduction to artificial intelligence to everyone interested as well as provide an impetus for exchange. In January, the MIT Technology Review, Süddeutsche Zeitung, Forbes and others reported on artificial intelligence.
Debating AI: IBM makes great progress
‘IBM’s debating AI just got a lot closer to being a useful tool‘ – MIT Technology Review, 24.01.2020
In June 2018, IBM presented one of its AI projects to the public under the title ‘Project Debater’: the AI software was developed to participate in a completely live debate with human experts and to produce convincing arguments in discussions. Now the MIT Technology Review reports that the so-called ‘Debating AI’ has made massive progress. This is because the software’s neural networks have been trained in recent months with some 400 million newspaper and research reports. Now Project Debater is able to deal with a wide range of topics and generate complex argumentation sequences on discussion topics such as ‘Should blood donation be made mandatory by law’?
More concrete plans: EU considers AI regulation
EU-Parlament will KI regulieren [EU Parliament wants to regulate AI] – Süddeutsche Zeitung, 20.1.2020
In a review also article worth reading, the Süddeutsche Zeitung outlines the status quo regarding the EU’s AI regulation plans, which are currently taking on a more concrete form. After a preliminary draft of a planned EU ‘white paper’ was leaked in mid-January – which, among other things, considers a ban on AI-based facial recognition – it is roughly known what form the planned regulations could take. On the one hand, the laws are to be all the more stringent the greater the potential for damage caused by AI applications. On the other hand, the legislative structure is to consist of a combination of so-called ‘horizontal’ regulations, which are to apply to all applications of artificial intelligence (liability issues, algorithm transparency, etc.), and of ‘vertical, targeted amendments’ to existing EU laws.
Managers want more commitment to AI implementation
‘Artificial Intelligence Can’t Get Here Soon Enough, Executives Say‘ – Forbes, 17.1.2020
A recently published study by management consultants KMPG concludes that managers are often frustrated by the slow pace of implementing AI technology in their own companies. According to the study, seven out of ten executives expressed the wish that their companies were more committed to the introduction of AI. A total of 751 executives from US companies were surveyed.
‘Bifold’: New AI Center in Berlin
Millionenförderung von Bund und Land für KI-Zentrum in Berlin [Federal and state funding of millions for AI centre in Berlin] – BMBF press release, 15.1.2020
The two existing AI competence centres of the TU Berlin – the Berlin Big Data Centre (BBDC) and the Berlin Centre for Machine Learning (BZML) – are merging at Salzufer in Berlin under the name ‘Berlin Institute for the Foundation of Learning and Data‘ (short: Bifold). Between 2019 and 2022, the Bifold is to be funded with more than 32 million euros from the federal budget. The state is also creating eight new AI professor positions at the TU Berlin and the FU Berlin. Anja Karliczek, Federal Minister of Education and Research, comments: ‘The merger of the two Berlin AI–competence centres into Bifold is a milestone in the implementation of the federal government’s AI strategy. Central AI competencies will be bundled in Bifold. Bifold will also focus on research into the transparency and comprehensibility of AI systems as well as ethical aspects‘.