Artificial intelligence is worse than humans in every way at summarising documents and might actually create additional work for people, a government trial of the technology has found.
Amazon conducted the test earlier this year for Australia’s corporate regulator the Securities and Investments Commission (ASIC) using submissions made to an inquiry. The outcome of the trial was revealed in an answer to a questions on notice at the Senate select committee on adopting artificial intelligence.
The test involved testing generative AI models before selecting one to ingest five submissions from a parliamentary inquiry into audit and consultancy firms. The most promising model, Meta’s open source model Llama2-70B, was prompted to summarise the submissions with a focus on ASIC mentions, recommendations, references to more regulation, and to include the page references and context.
Ten ASIC staff, of varying levels of seniority, were also given the same task with similar prompts. Then, a group of reviewers blindly assessed the summaries produced by both humans and AI for coherency, length, ASIC references, regulation references and for identifying recommendations. They were unaware that this exercise involved AI at all.
These reviewers overwhelmingly found that the human summaries beat out their AI competitors on every criteria and on every submission, scoring an 81% on an internal rubric compared with the machine’s 47%.
LLMs as they stand are already approaching the improvement flatline portion of the sigma curve due to marginal data requirements increasing exponentially.
It’s a known problem in the actual AI research field that nobody in private industry likes to talk about.
If it scores 40% this year it’ll marginally increase by 10% next year then 5% 3 years later and so on.
AI doesn’t follow Moore’s law.
So far “more data” has been the solution to most problems, but I don’t think we’re close to the limit of how much useful information can be learned from the data even if we’re close to the limit of how much data is available. Look at the AIs that can’t draw hands. There are already many pictures of hands from every angle in their training data. Maybe just having ten times as many pictures of hands would solve the problem, but I’m confident that if that was not possible then doing more with the existing pictures would also work.* Algorithm design just needs some time to catch up.
*I know that the data that is running out is text data. This is just an analogy.