All new technological developments require rules for use. As part of its digital strategy, the European Union has introduced the AI Act to regulate the use of artificial intelligence (AI). To the extent that robotics uses AI, the EU AI Act has implications for roboticists, too. In June, the European parliament voted to adopt the AI Act. As such it may well be the most far-reaching piece of legislation governing AI in the world so far.
It’s fair to say that 2023 is the year when AI entered the mainstream of public awareness. There’s no doubt that this year the public media – and not least parents – were first alerted to the power of so-called generative AI to help with homework and essay-writing. Newspaper stories about AI often focused on one particular system – ChatGPT. The main thrust of those initial texts was that generative AI would disrupt the testing of students by circumventing anti-plagiarism tools.
Concerns about AI’s potential impact quickly grew in the way that media scare stories often do. In a startling leap of scope, we were told that not only would AI do our children’s homework for them but it also posed a challenge to humanity itself. Soon, this sense of AI as a threat was further amplified: in March, prominent developers of AI called for a pause in AI’s advance so that society’s understanding of its potential – and therefore, how to regulate the use of AI – could catch up. Some observers sensed hysteria in the air, others recognised the risks but counselled caution, others still dismissed it as just another media fad. Roboticists will certainly recognise this spectrum of public reactions. This follows myriad examples of mass media misunderstandings and misrepresentation of robots and robotics. At the dawn of new technologies it was ever thus…
As the months have passed, the coverage has become more nuanced. The fear that fuelled those early headlines has been replaced by a more constructive sentiment about AI – opportunity. At this mid-year stage it is now easy to browse online and find dozens of articles offering advice on how to make the most of AI. In the field of education, reflective discussions about the implications (both pros and cons) of AI in the classroom and lecture hall can now be found online (as an example, see ‘Chat GPT (We need to talk)’ from the University of Cambridge). Indeed, in its announcement about the adoption by the European Parliament of its own AI legislation, the EU makes a point of emphasising the potential benefits, citing better healthcare, safer and cleaner transport, more efficient manufacturing and cheaper, more sustainable energy.
Of course, in the wider sweep of history, 2023 may prove to be insignificant – after all, who knows what will happen next year? Or maybe there was some new development last year or the year before that passed unnoticed at the time but that with hindsight may prove to be a major milestone. If future historians of AI do look back on 2023 as a significant year in the adoption of AI, it probably won’t be because this was the year that school students found a way to get their homework done without opening a book (whether paper or online). Instead, the milestone moment may well have been 14th June 2023. That was the date when the European Parliament adopted the EU’s first regulatory framework on artificial intelligence, known simply as the AI Act.
In adopting the legislation, the European Parliament set out its priorities: to make sure that AI systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly. Another key requirement is human oversight, that AI systems should be overseen by people rather than by automation, in order to prevent harmful outcomes. Following its adoption at EU-level, in line with the EU legal method, the AI Act is now being discussed at member state level. The current timetable is for the talks on the final form of the new law to be concluded and agreement reached by the end of this year.
At euRobotics, we’re keen to know what our members think about this major step towards the regulation of AI and, in particular, how they view its significance for robotics in general and for themselves in particular. If you have an opinion that you’re willing to share, let us know via firstname.lastname@example.org or leave a comment on Linkedin.