Uncle Ben, “With great power there must also come – great responsibility!” 
3 Likes
Hi Guys,
This thread has been “flagged” and as such I’ve been watching it since yesterday.
Whilst this thread doesn’t explicitly break any forum rules and is (IMO) actually quite interesting in a lot of ways it is also quite scary as - as Pete has already pointed out - the “review” has been assimilated into the general mire that is “the internet”.
I’m going to ask you all to take a step back and look at this thread as a whole and let me know what YOU feel should happen to / with this thread … I’m just curious as to your opinions. As it doesn’t break any forum rules (as far as I can see) then I can’t do anything with it without overstepping my boundaries but I will put my neck on the block and say that the thread in general leaves me very unsettled.
Phil
As a side note to all, you are no longer able to edit other people’s post titles. This was a default setting for the forum where users of a trust level of 3 could edit any post titles, but one could cause some real havoc with that.
You can still edit your own post titles, just not others.
3 Likes
Phil,
As I understood the initial post, using chat GPT to generate a review was just for fun !
I don’t think the credibility of that forum would be threaten by the saying of only one member, unless he is a c-level of dCS.
So, the only thing that disturbed me in this thread was when someone put a judgment, negative in that case, on the post of another forum member. This behaviour is very unusual on this forum, I do hope we will still be able in the future to post a comment without having someone judging it. Unless it breaks the forum rules, obviously.
What I like the most on the forum, is that dCS team members occasionally write some posts or answer forum members questions.
Like Ducati or BMW and I guess many other brands, dCS could organise a customer event…dCS meets Edward Green, would be a nice way to spend one day in England 
3 Likes
My 2c worth:
- My first reaction is that the “review” is indistinguishable from most of the drivel that many reviewers regularly trot out, a point I see Anup has already made. In fact one can but ponder how many “reviewers” are simply using ChatGPT to generate their reviews, or at least the first drafts thereof. I read an interesting article a while ago about AI-driven developments in quarterly results conference calls. AI tools are now used as a matter of course to analyse phraseology and vocal inflection for signs of negativity, particularly in unscripted elements such as answers to questions. Program trades are then auto-generated in small fractions of a second. The obvious way to preempt this is to have generative AI write the call script and deliver it. The (il)logical endpoint being that quarterly calls will very soon be presented by AI bots to AI bots, the results being analysed by further AI bots which will then generate trades by AI bots. This scares me more than deep fake Rossini reviews.
- As Anup also points out the Genie is now out of the bottle. LLMs now learn faster than humans and their rate of learning is accelerating. This raises an ocean of ethical questions, although in my opinion the integrity of the dCS owners’ forum is not high on that list.
- The strutslets provide me with some insights how society is dealing with this threat. My youngest (6th grade) is learning source criticism as a core part of social science, my middle (9th grade) is expected to apply rigorous source criticism according to a standardised model to all her cited sources in all essays in all subjects. My eldest (20 y o), now starting her second year of a law degree, has attended classes on AI ethics which interests her to the point that it may actually be something she pursues as a career path. Already as a second year she has been courted (lavishly) by McKinsey after she attended an seminar they sponsored on the subject of AI law that was actually aimed at final year students.
So were I land is: - This is out there. Just like guns, drugs and other potentially harmful human inventions it is incumbent on adult members of society to inform themselves about it and learn how to deal with it in a responsible way. Caveat lector.
- Generative AI has, within a very short space of time arguably turned source criticism into as essential a life skill as reading, writing or basic math.
- AI generated tomes on dCS products and Beethoven works are unlikely to lead to the end of the world.
2 Likes
chuckle I don’t think this was ever in doubt… 
I think my views are clear from the thread to date. So just a couple of points:
What rules? How do I find them? Other audio forums have the rules or link to them shown as a header. Others have the rules as a “sticky” at the top of the thread. I think this can be improved.
My second point is that the overall topic of AI and its use is a major issue of national and international governmental interest. The following links are to the current UK and EU positions. Of course it is not simple to correlate these to the particular issue at hand here. However regulation is certainly coming. Readers outside of these territories may well find equivalents elsewhere. In formulating future policy it may be appropriate to bear this and the underlying principles in mind.
Phil,
This is the message I get when clicking on the link you mentioned:
Oops! That page doesn’t exist or is private.
Thanks Chris - they should be published for all to see as far as I am aware so I’ll get that looked into.
That’s my goof - I posted my “admin” link to the forum guidelines that I can edit but you guys can’t see … the one for you guys to be able to read is here…
(I’ll also edit the link in my previous post…)
They can also be accessed by opening the “hamburger” menu at the top right and clicking FAQ.
2 Likes
@Phil Off topic, but while we are at it: the contact link at the bottom of this page actually is broken:
I have added an additional point to the section of the forum FAQ Be Who You Really Are:
The Discourse platform the forum is built on is very good at catching bots who create accounts and attempt to post on the forum automatically, and there are plenty of safeguards in place to help ensure posting is done by real people.
When it comes to content created by an AI platform however, It would be practically impossible for us to be able to police when such content is posted as, as far as I am aware, there is no tool available currently to detect it automatically. Instead, we have to rely on the integrity of the Community to notify other users when they are using content they did not create.
5 Likes
Updated, thanks for picking up on that & letting us know
2 Likes
Hi @Phil,
Thanks for picking this topic up. A couple thoughts:
I posted this AI-generated review with a clear title for the purpose of discussion. IMHO, there are many topics of interest that it raises, including those identified by @struts001, and @Anupc, and others.
I also felt the “review” to be a marvel of human achievement, since humans created ChatGPT.
I was surprised by @PAR’s reaction to my post, but Pete is a longtime member and contributor to the forum, and he is entitled to his opinion.
If, as you and Pete note–“the “review” has been assimilated into the general mire that is “the internet” and it is found and clicked on, the seeker would also find the post clearly labeled as generated by ChatGPT, and others discussing this fact. As such, I don’t see how that creates any problems, ethical or otherwise. Also, as noted, the review is quite positive for dCS ; )
I think it is not in the forum’s best interest to censor audio-related discussions, leaving its members wondering what they can debate and not. I also think that you, @James and others do an outstanding job in this capacity. For example, I have never seen you moderate discussions of competing products to dCS, which also shows dCS’s confidence in its own products, and in my case, strengthens my affinity for the brand.
Thank you for removing the capability to change titles. After @all2ofme 's post, I saw that I too had this ability, but stayed the hand of mischief. Your prompt action has stopped Ben from running amok.
; )
1 Like
There are lots (and while they are far from perfect they are in theory at least getting better at the same rate as the generators):
I’d consider putting ‘fake review’ in the title. While anyone here is bright enough to figure this out, it might get picked up by the search engines.
I have no problem with the topic as a whole, it’s an important subject to debate.
2 Likes
I have the Rossini APEX + Clock and the AI generated article is very accurate and “holy $hit scary” at the same time.
What a cool time to be alive.
Thanks for the enlightenment of how far machine intelligence has become.
I think you are OK without having to add “fake review” as you already have “AI generated” in the title. FWIW
Robots certainly have come a long way since “Danger Danger Will Robinson”
1 Like
I assume anyone considering a dCS product is doing quite a bit of research (and listening) before purchasing. I assume they read lots of reviews, and perhaps have review sources they already know and trust.
I doubt anyone would purchase a Rossini based on this AI generated review (though, not a bad outcome for anyone). And the AI review has a whole thread after it, so it’s obvious what it actually is.
The review reads like a lot of AI generated content these days – it’s fairly bland and lacks personality. I was talking with a neighbor who teaches college level and has seen AI generated essays submitted by students; he can identify them because of their blandness, genericness. (The challenge is how to call the student out on this and what to do about it, because you cannot absolutely confirm it was AI generated.)
My son is applying to college and writing a lot of essays. He says he’s not using ChatGPT, and I trust he isn’t, but in the end who knows. With acceptance rates as low as 5% you really have to stand out from the crowd, and I don’t believe an AI generated essay will help you do that. The essay has to be well written and come from you and your unique experience, ideas, POV.
As to what to do with this thread, I leave that to dCS to determine. It’s a fake review, but the thread that follows is interesting, timely and valuable.
Many companies and industries are grappling with similar questions these days. Many new lessons for all of us to learn.
5 Likes
AI lacks the capability to listen and form its own opinions when evaluating a Hifi product. Writing a review differs fundamentally from creating a fictional story. In such reviews, all impressions and subjective assessments are fabricated, potentially leading to misinformation. This issue is compounded by search engine indexing, as it may lead casual searchers to mistakenly assume the review is from a real user.
However, when a post is clearly labeled as AI-generated and meant for discussion rather than deception, there should be no cause for concern. By openly discussing the ethical implications and advocating for responsible handling of AI-generated content, we contribute to the broader dialogue on the ethical use of AI.
7 Likes