You've reached our premium or archival content
To access this page, and more great content just like it, you need to become a paid subscriber.
If you already have an account, please login.
Otherwise, registration is quick and you'll have access instantly after payment.
I have made comments about AI applied in broadcasting.
These bots may well give the listener the feeling that they are listening to a real person.
These bots "scrape" the web for information and AI models are trained on the information.
Bots are a great research tool but should be used judiciously.
From my experience, I tried chatGPT by entering my advanced contracts essay I attempted in 2010 and asked a programming question.
When I entered my advanced contracts essay that I attempted in 2010 into the chatGPT program.
The answer produced by chatGPT was totally irrelevant to the topic and lacked references.
When I attempted the unit in 2010, I was awarded an HD.
If I marked that essay produced by chatGPT I would have awarded it a FAIL.
Then I was asking chatGPT to program and process data in an array. The answer produced a basic error in array processing by starting indexing fron 1. Indexing arrays start at 0.
I told the bot that it was wrong and the bot "apologized" and produced the correct output with indexing starting at 0.
On another occasion I asked about Cardinal Pell (RIP) and the output from chatGPT said that he died in 2021 when in fact he died in 2023. I told the bot of that error and again it apologized.
So far we have to be careful of taking output from chatGPT as 'gospel'.
In another post on this site about bots as synthethetic presenters how management has to be careful about content produced by a bot which may be prejudicial and/or perjorative to people due to race, disability and sexual preference.
Consequently, as a result of a bot producing prejudicial and perjorative content, management may be subject to attending state and federal anti-discrimination and human rights tribunals. Maybe management may have to justify holding their licence before the ACMA.
Another issue is how nuanced is the output produced by a bot. The article does mention that bots can be expressive and conduct interviews. How expressive is that over listening to a orogram over a long period. It may well get boring.
As I said in a comment in a similar post, these bots will lack the sponteneity of a Ward Austin (RIP) or Doug Mulray.
I would fail to hear how a bot can do a talk show like John Laws or Ray Hadley or The Health Report or call a football match.
Bots may be ok for shows with little or no commentary or short news bulletins. However not for long talk shows.
Then if all the radio stations in a particular market use bots, I predict that the shows across all stations will sound the same with stereotypical lines and slogans going to air.
Bots will only be useful for web content that is static but may not be able to capture listeners' changing views.
Bots will definitely not be able to innovate new ideas in programming where the information on the web is scant but not available.
Thus a bot may be useful as a research tool and for links to news items and music but not for talk shows.
That provided that the content makes sense, is relevant and is not prejudicial.
Thank you,
Anthony of be judicious Belfield in the land of the Wangal and Darug Peoples of the Eora Mation