Companies worldwide are beginning to use LAMs or Large Action Models, according to Adnan Masood, UST’s Chief AI Architect. LAMs are advanced artificial intelligence systems that can understand natural language and act to complete tasks on their own. They’re more powerful than ChatGPT-like systems because they interact with various applications and devices to perform real-world jobs, said Mr. Masood. He added that while planning a vacation or optimizing an investment portfolio, they also learn and adapt based on user requirements.
Saving labor costs
Labor cost reduction is among the many benefits these models bring to companies. To give a few examples; American insurance agencies automate claims processing with LAMs; customer service at a major European airline has been improved thanks to them; retailers throughout Asia have seen a 25% boost in sales through personalized recommendations made by LAMs.
LAMs can process different types of inputs — speech, video, text — and navigate through user interfaces just like humans do using this information coupled with APIs & IoT devices which makes them extremely versatile according to Mr Masood.
One early demonstration of this technology was done at CES (Consumer Electronics Show) – where a device powered by LAM showed seamless handling of multiple tasks. It is expected that businesses will either build their own LAMs or buy them from providers offering such services as software services or cloud computing suggested by Mr Masood who believes that eventually every company will have one.
Impact on Jobs
When asked about possible job threats posed by LAMs, Mr Masood admitted that yes indeed some tasks currently performed by people could be replaced but new employment opportunities will arise as well since these machines enable novel functionalities while freeing up humans’ time for creative higher-level work.
Training oversight roles would need creating so too would explanation giving around actions taken by such models but transparency fairness security must never be compromised during development of large action models thus there should always be tests for biases and misuse coupled with clear responsibility also proactive reskilling will be required to manage changing job types brought about by LAMs in addition governments need policies supporting upskilling suggested by Mr Masood.
Source: thehindu
