Should we believe in the hype of generative AI?

Should we believe in the hype of generative AI?

Technology advancements are accelerating at an exponential rate. At the forefront of that is the rapidly evolving world of artificial intelligence. Is the hype all it’s made up to be? Or should we be approaching with a more cautious view?
Author: Ark 51 – ark-51.com

 

Technology advancements are accelerating at an exponential rate and at the forefront of that is the rapidly growing, improving and evolving world of artificial intelligence (AI). Is the hype all it’s made up to be? Or should we be approaching with a more cautious view?

Generative AI and Large Language Models

 

Generative AI typically takes the form of large language models, with the most known one being Open AI’s Chat GPT models. These models use very complicated calculation algorithms to effectively “guess” the next best thing to say, given the context of the question asked and the words the answer has generated up to that point. To the human brain, it appears to be thinking, contextualising and answering the question put to it in a very humanistic way.

 

On the face of it, this is a brilliant advancement and can be used in a limitless number of optimisations for a company. Can I ask AI to draft me a report with the data I have? Can it summarise the information in a document? Could it extract key terms? Even more complicated processes could, in theory be streamlined and made more cost effective by “throwing AI at it”.

 

Where should we be cautious?

 

Generative AI used for its primary purpose of text generation – which includes data extraction, summarisation, or any other text transformation task – can result in very cost effective, transformative changes to a business’s processes. The issue comes primarily from shareholders and decision makers not understanding the limitations of these models, which is driven primarily through the medium in which they are built – probabilistic text generation. This means that they have not got a capacity for anything requiring reasoning, or higher order thinking. Asking an AI where to allocate collateral in financial contracts, or similar, will result in sub optimal results because the process of allocating that resource requires a much higher amount of reasoning and ‘thought processing’. Something they intrinsically (in the case of Gen AI and LLMs) are not designed to do.

 

Should we believe the hype?

 

Absolutely. The age of AI accelerated technology is already here. But believing it is a one solution to fit all problems tool is the potential downfall. The hype is fully worth believing in if the problem faced is one that text generation would make a significant impact on. The most obvious and potent one that businesses can utilise is to extract the data already stored in PDF and word document files and converting them into manipulatable, analysable data in a fraction of the time taken without using AI.

 

Ark51 is a platform that already does this. Combining the subject matter expertise of legal contracts from our team of in-house legal consultants and solicitors with cutting edge AI, the SAAS solution can quickly ingest, parse and provide clients with macro and micro views of data, along with bespoke reporting, analytics and workflow considerations. The application can supercharge handling data and give an edge in a very competitive industry. If you would like to see a demo, or talk with someone about how we can save you time and money, please reach out to [email protected]

Make possibility reality

Become an IA FinTech Member
and see where it takes you.

Open-Lock_icon.png
Login to your account