Are we now in an era of just AI talking to AI?

Ozzy

Founder of UKBF
UKBF Staff
  • Business Listing
    Feb 9, 2003
    8,497
    11
    3,595
    Northampton, UK
    richardosborne.co.uk
    I've just come from another thread that started as one question, but then moved into discussing AI replies, and I'm seeing this more and more in this forum, I am also seeing it more and more in emails I received.

    I even find it now in my email client, I click to reply to someone and before I have a chance to type a reply Google AI has already checked my diary and typed as a reply saying "Yes I can make 10am on Wednesday, please send an invite and lets chat then"... and I'm like yeah that works and just click Send....AI has replied to an email to a work colleague for me and all I did is click Send.

    People use AI on this forum, sometimes to help articulate something better, sometimes to help when English is a second language, and sometimes to spam the forums and post junk. As moderators, that does make it tricky being pragmatic when AI is merely used as a supporting tool.

    But with AI tools exploding in every part of our lives, are we now simply the flesh bags carrying the shopping in as AI now chats to AI on our behalf in work and home?

    Aside to that, using my email reply above (that really did just happen) is that a good thing? It did save me a little time and if that happens a lot that saves me a lot of time.
     
    I mentioned in another thread. I use voice dictation heavily in my workflows now as it enables me to think naturally, talk clearly and use AI to help synthesis ideas into writing.

    This workflow has massively increased the amount I can contribute on forums, email, slack and other channels to help people and share knowledge. I've also trained my AI on my writing, so it should use my personal tone of voice and style. Also I'm not a natural writer, writing things by hand takes me a long time and I find myself overthinking and over editing. So yes, i think AI as part of a human workflow can be incredibly helpful. But you can occasionally see the AI artefacts like em-dashes and paragraph structures in its writing.

    I think this will become the norm for many - human driven AI assisted writing. It could increase the bar and quality of human writing to be clearer, more concise, more accurate if humans manage the output well and provide consistent regular feedback. I am supportive if it broadens the contributions and enables more humans to contribute their knowledge faster.

    (PS I wrote this by hand, so as not to offend anyone, but you can see it is rambling, and I took too long overthinking it and re-editing it! Excuse the spelling errors)
     
    Upvote 0
    I think this will become the norm for many - human driven AI assisted writing.
    I agree, it will become the norm if it is not already nearly there.

    I used Claude consistently for much of the day in many aspects of my work. I offload my ideas and thoughts into it, point it in the direction I need or want some further research, and then ask it to collate into a report for myself and others to digest.

    I ask it to review discussions on this site, I add my thoughts on the discussion or the subject matter, and I feed it with conflicting perspectives and theories, and then ask it to draft an article for consideration alongside asking it to read the thread and pull out members comments ensuring different perspectives are shared. Probably saves me about a full day of time.

    I've also just last week submitted my response to a public consultation on Alternative Provision and funding for children with special needs (that was quietly being crept out under publicised). Claude was able to review various legislative documents, and parliamentary debates for me in a fraction of the time it would have taken me. I gave it my stance on various questions being asked on the consultation and it was able to articulate those and write the consultation response for me. Probably saved me about a week of time!

    So personally I am a fan of AI, but only when used constructively or supportingly. Used for dancing cat videos, or just used for writing some stuff with zero input by the 'human' is where it is a waste in my opinion.
    Or like me saying "write an article/forum post on mtd' for posting on here, lazy and no creativity input from myself or taking into account the views or advice of the members here.
     
    • Like
    Reactions: AI THINK TANK
    Upvote 0
    I'm using AI more and more for coding (because I'm lazy). But for writing I always begin with putting down the framework by hand. I fill in and flesh out then give it to an AI tool to refine and adjust for my chosen audience.

    As an example I'm the Chair of a board of Trustees and need to do the annual report. I wrote the report but AI made is more 'official'. I wrote a summary of an event and AI made is more friendly.

    But when someone gives an AI tool a prompt and says 'write me an X about Y' it just doesn't read quite right. The sentence structures are inconsistent or the style is off. It just feels wrong. You can usually spot the AI posts here on the forums.
     
    • Like
    Reactions: Ozzy
    Upvote 0
    I'm using AI more and more for coding (because I'm lazy). But for writing I always begin with putting down the framework by hand. I fill in and flesh out then give it to an AI tool to refine and adjust for my chosen audience.

    As an example I'm the Chair of a board of Trustees and need to do the annual report. I wrote the report but AI made is more 'official'. I wrote a summary of an event and AI made is more friendly.

    But when someone gives an AI tool a prompt and says 'write me an X about Y' it just doesn't read quite right. The sentence structures are inconsistent or the style is off. It just feels wrong. You can usually spot the AI posts here on the forums.
    I am afraid I cannot tell AI posts, possibly because I am unaware of using it myself (I realise it may get itself involved in some of my work). But I think we need to be very aware, on here, of the value of AI for people whose written communication skills vary. And there is a need to be very careful not to attack people because their communication style is not the same as some of us may be used to.
     
    • Like
    Reactions: Ozzy
    Upvote 0
    Did any of you ever watch the movie "Mortal Kombat: Annihilation" where Raiden criticises Jax for relying on tech instead of his own inner human strength? Its kind of analogous to Generative AI augmenting people with powers to write, code, provide assistance and so on.

    Human language is for humans (yeah alright, also for shouting at pets, monitors, cars, smart assistants etc.) and we are technically legal entities, which means we own what we say even if its generated by an LLM or another person. The idea of "AI talking to AI" but using the human as a medium wouldn't change the ownership of what was said. Last time I checked, an AI/LLM cannot be sued for something that could be deemed offensive. Not sure about the company/person who's trained that particular model.

    For "AI talking to AI" directly without the human as the medium, I don't think they'll need the whole NLP (natural lang processor) layer, but whether they can be appointed to legally speak on behalf of a human or a company, I'm not sure if we're there yet, but I've heard of an AI CEO, but I don't think it had full authority as that would open up legal questions.

    Another point, software languages exist to enable humans to instruct computers. At some point, high-level programming languages won't be needed as future AI agents will likely build machine code directly, skipping the need for heavy compiling or human-abstracted languages entirely.

    What will happen to us bags of 80% water or "flesh bags" as Ozzy put it. I don't believe there's a single answer but I reckon we'll come full circle at some point, just like Gen Z ("Zed"!!) are supposedly ditching smartphones for basic brick phones just to make calls and send SMS.

    (Sorry... that was a bit of a ramble... but my excuse is, I'm multi-tasking)
     
    Upvote 0
    The true purpose of AI is to allow wealth to access skill without allowing skill to access wealth.

    On paper AI is supposed to help humanity, but really it’s just another commercial engine. We should also keep in mind AI has been around for decades, supporting hard-core critical industries. The problem we have these days is that a 70-year old engineering discipline is always being reduced to modern corporate conspiracy, completely ignoring history.
     
    Upvote 0
    But I think we need to be very aware, on here, of the value of AI for people whose written communication skills vary. And there is a need to be very careful not to attack people because their communication style is not the same as some of us may be used to.
    Tools to check grammar and spelling have been around for a long time. And my spolling and gramma tends to creep into my posts. I'm always spotting typos. But getting AI to write the whole post isn't the answer. I don't mind a poorly written post as long as it makes sense and adds value.

    Too many AI written posts rabbit on and on and don't offer anything of value. On another forum I used to frequent they employed an AI bot to answer posts. It was almost always wrong and you did end up with bots answering bots.

    So while we might end up with more bot written threads/post I suspect human readership will diminish as it becomes less and less relevant.
     
    • Like
    Reactions: EnterprisePro
    Upvote 0
    AI has been around for decades, supporting hard-core critical industries.
    Disagree with that statement. We have had ML for decades. And much of what is now called AI is still ML. The AI we all use is really just a stochastic parrot.
     
    Upvote 0
    ok.. I hear you, however, just to pick out one industry that I'm closely familiar with is aviation. Aerospace firms have implemented expert systems and artificial neural networks in flight control software for a very long time. Autopilot, automated landing systems and collision avoidance protocols (TCAS) are deterministic and predictive AI systems. Human pilots have acted as safety monitors for an AI doing the heavy lifting. "AI" is an overarching umbrella term for the entire field of making machines mimic intelligent behaviour. AI was in fact part of my degree, and this was around the time Google was just a server box in Stanford Uni. The Dr. who lectured us on AI principles covered the use-cases in industry, as well as good old chess.
     
    Upvote 0
    Where did he park his Tardis? 😁😁

    It's not a bad question! Not nice for me to be stereotypical, but the fella was, different, shall we say. He was always in an anorak, but being a consultant in the defence industry he was (seemingly for us young immature students) quite a big deal from our perspective.
     
    Upvote 0
    That may be for modern aircraft but for things I worked on back the 80s it was at best very basic ML. The words AI didn’t even exist (except in the books Asimov wrote).

    The problem we have is AI is used everywhere without good reason. The LLM used by AI to write forum posts doesn’t learn, it doesn’t understand or have memory. The models used to predict the weather are better but still just ML.
     
    Upvote 0
    while we might end up with more bot written threads/post
    This would only be the case in my worse case scenario from my OP, where AI written posts are generally the only part of the post. In my opinion using AI tools to help articulate isn't an issue as the human has something to say, a question to ask or answer, and just wants some help to get their own genuine message across.

    Then there's the other part of my OP, where AI built into Gmail or Outlook offers up suggested replies to emails. Which then receives another AI suggested reply, and the cycle continues 😁. That is where we have AI talking to AI.
     
    • Like
    Reactions: fisicx
    Upvote 0
    I use Thunderbird which has remained AI free. And Libre Office also AI free.

    My biggest issue with AI (other than the spyware) is environmental cost. The industry is trying to greenwash their impact but it’s not working or sustainable.
     
    Upvote 0
    I've just come from another thread that started as one question, but then moved into discussing AI replies, and I'm seeing this more and more in this forum, I am also seeing it more and more in emails I received.

    I even find it now in my email client, I click to reply to someone and before I have a chance to type a reply Google AI has already checked my diary and typed as a reply saying "Yes I can make 10am on Wednesday, please send an invite and lets chat then"... and I'm like yeah that works and just click Send....AI has replied to an email to a work colleague for me and all I did is click Send.

    People use AI on this forum, sometimes to help articulate something better, sometimes to help when English is a second language, and sometimes to spam the forums and post junk. As moderators, that does make it tricky being pragmatic when AI is merely used as a supporting tool.

    But with AI tools exploding in every part of our lives, are we now simply the flesh bags carrying the shopping in as AI now chats to AI on our behalf in work and home?

    Aside to that, using my email reply above (that really did just happen) is that a good thing? It did save me a little time and if that happens a lot that saves me a lot of time.
    Apart from helping society with for example medical issues and other research, we should STOP using AI. It takes huge amounts of water to cool the servers.
     
    Upvote 0
    The problem is that alternatives would not be commercially viable.
    The AI industry is not commercially viable. Vast amounts of capital investment, off balance-sheet loans, special purpose vehicles, circular investments, and users only charged a fraction of what it costs.


    Aside to that, using my email reply above (that really did just happen) is that a good thing? It did save me a little time and if that happens a lot that saves me a lot of time.
    I try to avoid that as much as possible. Maybe it saves time now, but at the expense of losing our ability to form sentences and putting the effort into making our thoughts understandable.
     
    Upvote 0
    Bands and events seem to be using AI now for their adverts, which apart from anything else, makes them all look the same!
     
    Upvote 0
    I've just come from another thread that started as one question, but then moved into discussing AI replies, and I'm seeing this more and more in this forum, I am also seeing it more and more in emails I received.

    I even find it now in my email client, I click to reply to someone and before I have a chance to type a reply Google AI has already checked my diary and typed as a reply saying "Yes I can make 10am on Wednesday, please send an invite and lets chat then"... and I'm like yeah that works and just click Send....AI has replied to an email to a work colleague for me and all I did is click Send.

    People use AI on this forum, sometimes to help articulate something better, sometimes to help when English is a second language, and sometimes to spam the forums and post junk. As moderators, that does make it tricky being pragmatic when AI is merely used as a supporting tool.

    But with AI tools exploding in every part of our lives, are we now simply the flesh bags carrying the shopping in as AI now chats to AI on our behalf in work and home?

    Aside to that, using my email reply above (that really did just happen) is that a good thing? It did save me a little time and if that happens a lot that saves me a lot of time.
    I think we're definitely entering an era where AI is talking to AI more often than most people realise. Your email example is probably one of the less controversial uses of it.

    But, at some point you have to ask whether the humans are still communicating or just approving machine-generated intent.

    We've found that's exactly why there's so much interest in AI speakers and experts at the moment. The technology is moving so quickly that most business owners aren't struggling to find AI tools, they're struggling to work out which changes actually matter to their business.

    Speakers like Cassie Kozyrkov don't focus too much on the latest shiny tool. Instead, she talks a lot about decision-making, where AI fits into an organisation and how leaders can use it to remove low-value work while keeping humans responsible for the important calls. That feels far more useful than simply chasing the newest feature.

    I suspect over the next few years the businesses that get the most from AI won't necessarily be the ones using it everywhere. They'll be the ones that understand where it genuinely improves outcomes and where people still need to be firmly in the loop.
     
    Upvote 0
    I suspect over the next few years the businesses that get the most from AI won't necessarily be the ones using it everywhere. They'll be the ones that understand where it genuinely improves outcomes and where people still need to be firmly in the loop.
    It's more likely to be those that can afford to use AI. Price gouging is just beginning.
     
    Upvote 0
    I expect like all technical innovations, they are used in the main, not to reduce the burden of working at mundane tasks, but to increase profits and reducing numbers of staff who are often made redundant.
    I for one still will not use self checkouts.
     
    • Like
    Reactions: gpietersz
    Upvote 0
    most business owners aren't struggling to find AI tools, they're struggling to work out which changes actually matter to their business

    Yet another unfounded claim (thanks to AI slop).

    Which research institute published this, which business segment did they study, and in what geographic location?
     
    Upvote 0
    I think the useful distinction is whether AI is augmenting someone's thinking or replacing it. If it saves you from repetitive work or helps turn your own ideas into something clearer, that's a productivity gain. The problem is when people outsource the thinking entirely and just post whatever the model produces.
     
    Upvote 0

    Latest Articles