Technology and Innovation Community

 View Only

  • 1.  AI & Policy advice

    Posted 29-04-2025 14:46
      |   view attached

    Hello community,

    I came across this article in New Scientist (March 2025) - attached. New Scientist obtained the ChatGPT records of the UK's technology secretary Peter Kyle, using freedom of information laws. Tech Secretary asked this AI about UK's small and medium business community being slow to adopt AI, ideas on his media appearances, podcasts etc.

    The article quotes an opinion of John Slater, an FOI expert, who says that ChatGPT and Google are not the same, ChatGPT does "create" something new-ish based on the input from the users". With this uncertainty, politicians might want to avoid using privately developed commercial AI tools like ChatGPT", said Jon Baines, lawyer, Mishcon de Reya.

    How this impacts what investors do when they use LLMs? Is this a "can of worms" kind of question for us?



    ------------------------------
    Aya Pariy
    ------------------------------


  • 2.  RE: AI & Policy advice

    Posted 29-04-2025 17:26

    Hi Aya

    I think the main issue here is the lack of human supervision of the LLM output. The investment industry must embrace LLMs without taking for granted that they can replace humans without proper human  supervision. Although models' hallucination issues will continue to improve in the future, the rationale of the output can still be judged by a human supervisor, especially in critical decisions like nationwide policies or critical investment decisions.  

    Happy to know the opinion of other community members.



    ------------------------------
    Carlos Salas
    Portfolio Manager & Freelance Investment Research Consultant
    ------------------------------

    Original Message


  • 3.  RE: AI & Policy advice

    Posted 30-04-2025 10:52
    Edited by Todor Kostov 30-04-2025 18:27

    Hi Aya,

    Thanks for the information and the various feedback from the experts.

    I agree with Carlos that the final decisions based on usage of AI tools/ LLMs (regardless of whether they are developed from commercial or public organisations) should be taken by human beings. Same applies to oversight/supervision of tasks on model outputs.

    Recent example is the insight by David Solomon, CEO of Goldman Sachs, at the Cisco AI  Summit in January, revealing that already AI can draft 95% of an IPO prospectus in minutes but that "the last 5% now matters because the rest is now a commodity". 

    Happy to hear what others think or have already experienced in their line of work.

    Todor



    ------------------------------
    Todor Kostov
    Director
    ------------------------------

    Original Message


Related Content