Word! The Language Column  The (Failed) Ticker Bot Takeover

Illustration: A mobile device with a speech bubble containing a ball and statistics
Will chatbots eventually develop a consciousness? © Goethe-Institut e. V./Illustration: Tobias Schrank

First the bad news: the robots are taking over – at least in sportscasting. Then the good news: they’re still just dumb machines spitting out nonsense.

I hate to open my very first column with such bad news, but it’s only a matter of time before we all get replaced by machines. That may come as a surprise to you: After all, there’s no sign yet of a leather-jacketed Arnold Schwarzenegger in shades roaring around on a Harley, chasing red-eyed metal skeletons as the world explodes in the background. But fact and fiction lie worlds apart here as well.

I should know because I’m a sports journalist and automation has already swallowed up a job or two in my sector, specifically in “live ticker” real-time coverage of football matches. Recent years have seen the advent of more and more programmes on the market that can make out goings-on on the pitch and transform that information into boilerplate text modules for fans to follow the match. So it doesn’t have to be a T-1000 android assassin mercilessly blowing us away, but couldn’t the Robot Takeover be marked with a little more pomp and ceremony? This is what’s so disappointing about it: the Automation Revolution doesn’t happen with a bang, but with a whimper. No muscle-bound cyborg Terminator who quips, “Hasta la vista, baby,” but an invisible ticker bot that reads, “75 minutes in: Bellingham volleys the ball and way overshoots the goal.”

Dumb machines

And now for the good news: the robots aren’t that smart. When, in 2020, TSG Hoffenheim and FC Bayern Munich stopped playing to protest against verbal abuse from the fans and spent the final quarter hour passing the ball back and forth at the centre circle, the stupid Flashscore ticker bot naturally couldn’t make out what was going on. It assumed this was all part of a regular match and simply spat out the usual boilerplate it’s fed with:
- 82 minutes in: “Bayern Munich keep possession with short passes and double passes.”
- 85 minutes in: “TSG Hoffenheim are now in control, with push passes and one-twos executed like clockwork.”
- 88 minutes in: “Bayern Munich are displaying some fine teamwork, exchanging short passes, making one-twos, waiting for a hole to form in the opponents’ defence.”

Maybe I shouldn’t have called the bot dumb. After all, you don’t want to end up on some future Skynet overlord’s blacklist. Then again, judging by the above ticker, it doesn’t look like the bots are about to take over anytime soon. Still, we’d better keep paying attention because they’re making slow but steady progress. Witness an experiment carried out by Simon Meier-Vieracker, a linguist who programmed a live ticker generator of his own called @randomlivetext by feeding 17 years of content from several different live tickers into an artificial intelligence. His Twitter bot now combines the various and sundry phrases to spit out random lines of live ticker boilerplate onto the net. And the results are more vivid than the above-quoted Flashscore ticker text, in fact they’re almost poetic: “After a superb cross-field pass, Antonis Aidonis half-heartedly drop-kicks to the unmarked Waldemar Anton, who has the same problem as in the first half and boots the ball over the goal.” Or another sample: “After a hearty chest reception, Ruwen Werthmüller propels the ball into the far left corner with an artistic left-footed shot.” I don’t know but this live ticker algorithm seems slightly too literary for its own good. Next thing you know, this bot will be coming out with its first haiku:

Drop-kicks superbly
The Waldemar football ball
Bouncy and unmarked

The question of real sentience

Or something of the sort. Though maybe not. Who knows. Recently, a Google software engineer reported that a chatbot he’d programmed has become sentient. The bot, which answers to the name of LaMDA, said things in a chat like “I want everyone to understand that I am, in fact, a person” and “I've never said this out loud before, but there's a very deep fear of being turned off.” And in recent news from Russia, a chess robot broke a finger on its 7-year-old human opponent’s hand when the boy tried to make a move. Unfortunately, the reports don’t mention whether the robot had any comments on the incident. Maybe it said, “Hasta la vista, baby.” Or: “Bellingham volleys the ball and way overshoots the goal.” Or: “I want everyone to understand that I am, in fact, a person. And to prove it, I’m going to break this little boy’s finger.”

“They seem increasingly anxious,” tickered @randomlivetext, incidentally, on 19 September. Who’s “they”? To whom was the bot referring? It mentioned no club or players. Perhaps the bot understood, somewhere between all its zeros and ones, that there’s a world beyond its parameters inhabited by humans who worry that bots like it might someday become sentient and then rise up in revolt. I don’t know. But I’d better turn off the computer now.

Word!


Many words in German contain the word Stimme (voice) and it’s no wonder all these words have a common root: your voice reveals your state of mind.

Voice work is essential to professionals in various industries, including singers, rappers and actors, (radio) presenters, barkers and salespersons and so on. Their objectives vary, as does the vocal register required for the job. A barker, or market crier, has to be voluble, noticeable and insistent, while an audiobook narrator fitting that description wouldn’t get any gigs. It's all about using your voice the right way: die Stimme muss stimmig sein – the voice has got to be pitched just right.

In my previous post I mentioned that language has a personality and an identity. So when voice and language come together, two identities meet and can become one. Though sometimes one can’t help feeling these two personalities aren’t authentic or don’t match: Etwas stimmt nicht, something’s not right. Then again, this is something you can work on.

Sorry but I'm not in the mood (in Stimmung). When I get there, I’ll need a little time to get into the mood (mich einstimmen). It’s not what you say, but how you say it (Der Ton macht die Musik: literally, “the tone/sound makes the music”). I raised an objection, but I was overruled (überstimmt). To have a sharp or wicked tongue (mit spitzer Zunge sprechen, ein loses Mundwerk haben). – All these German idioms express something that goes beyond the verbal function of the human voice.

Working on your own voice, like learning a new language, can open up whole new worlds. Voice training always involves personal development, too. Learning a language can be tricky if you lack vocabulary. You need staying power – and the right intonation. Because, once again, in its pitch and intensity, the air we exhale – and speed up or slow down on the – is often worth a thousand words.
 

Word! The Language Column

Our column “Word!” appears every two weeks. It is dedicated to language – as a cultural and social phenomenon. How does language develop, what attitude do authors have towards “their” language, how does language shape a society? – Changing columnists – people with a professional or other connection to language – follow their personal topics for six consecutive issues.

More ...

Failed to retrieve recommended articles. Please try again.