CHOICE #1 WIND DOWN HUMAN INTERPRETING My eyes sting a bit as I write this. The first and easiest choice is to accept that AI is inevitable, you cannot put the toothpaste back in the tube, cannot turn back the tide. All those clichés mean accepting of AI as a replacement for human interpreting. In that thinking, we have 5, maybe 10 years left. In some ways, this is a tempting route. It takes little to no effort, allows us to pile pressure on freelancers to adapt and diversify, and sounds valiant. We can do the honourable thing and give up while we are ahead. Let’s swipe a few errant pastries on the way out and will the last human interpreter please turn out the lights? CHOICE #2 COLLABORATE WITH THE MACHINES This choice has become more popular nowadays. It suggests that the way to keep human interpreting is to adapt the technology for our own purposes. We can use Computer Aided Interpreting tools, lean on Automated Speech Recognition and mine terminol- ogy with AI and summarising tools. That all sounds great and honestly, some of the research looks promising. We can’t ignore the fact that those tools make us more precise with our ter- minology, names, and dates and that, if used well, some might speed up preparation. It sounds incredibly tempting to try to co-opt tech- nology to our side, rather than having it only in the hands of big tech companies. CHOICE #3 RETHINK INTERPRETING This is the most difficult choice of all. Some people, looking at how AI and technology work, have been arguing that interpreting needs nothing more than a complete rethink. Remote interpreting makes the concept of a professional domicile tricky. Machine interpreting makes the “don’t add or omit” ap- proach seem short-sighted. Tech companies selling dreams of futures with no language barriers make marketing on the basis of accuracy, neutrality, and impartiality seem about as useful as a deck chair on the Titanic. This option means being ready to critically reas- sess how we talk about interpreting, the expectations ECHOS AUS DER BERUFSWELT Jonathan Downie ist Kon fe renz- dolmetscher, Übersetzer, Redner und Autor und lebt in Edinburgh, Schottland. Er interessiert sich für die Verbindung von Forschung und Dolmetschpraxis sowie das Dolmetschen in Kirchen. Sein erstes Buch, „Being a Successful Interpreter: Adding Value and Delivering Excellence“, ist im Mai 2016 erschienen. Seine Kolumne zum Thema Forschung als Mehr wert für Dolmetschkund:innen erscheint exklusiv im VKD-Kurier. Jonathan Downie we have of interpreters, and even how interpreters work, to make sure that interpreters are viewed as experts doing work computers can’t replace, rather than automatons who turn pastry and coffee into words and invoices. NO CHOICE IS PERFECT The problem with Choice #1 is, of course, that all the available data are telling us that machine interpret- ing is nowhere near replacing us and still produces errors that create real risks to people’s reputation and future. The vast majority of conferences should still be using humans and giving in to the wave of AI interpreting means ignoring the problems with the technology. That is hardly professional. Choice #2 is not wrong in itself but it is probably wrong to see it as a way of saving human interpret- ing. After all, human translators adopted Computer Aided Translation, automated Quality Assurance, fancy glossaries, and a whole load of other tech- nologies and yet still translator pay is down, agency profitability is down and, according to the ELIS 2026 survey, everyone in the sector knows that quality is down too. There is simply no evidence that techno- logically augmented interpreters would be any bet- ter placed to keep their jobs than interpreters who still turn up with nothing but a notepad and a folder full of paper. Choice #3 means upsetting people. It means hav- ing successes and failures. It means learning to com- municate in new ways and thinking about interpret- ing in new ways. It means constantly being out of our comfort zone and regularly facing down opposition. Rethinking interpreting is never easy and requires us to accept when we are wrong or when what we pre- viously thought no longer works. JUNI 2026 VKD-KURIER 11