GPT-3 has been hailed as an artificial intelligence breakthrough. John Thornhill tries it out and assesses the rewards — and the risks.
It is easy to forget that AI has been with us for a long time. Banking and Finance leverage AI back in the 1950's.
This revealing article is an aid for all executives struggling to embrace AI into long term vision, goals and strategies.
It remains the case that AI fits narrow use cases which exhibit predictability and therefore help in automation and freeing up people from these low value-add but cumulatively important tasks. Free up time and creativity for the more complex high value-add tasks that require human intuition, empathy and creativity.
Those working to leverage GPT-3 are the first to highlight that even this most advanced application of AI is "capable of stringing together plausible-sounding sequences of words without having any concept of understanding. As Turing foresaw decades ago, computers can achieve competence in many fields without ever acquiring comprehension.
That is why automation and straight-through-processing (STP) are alluring but not a panacea to cut costs and increase efficiency. In many cases plain rules based engines are as effective as the best algorithms so a pragmatic approach is necessary using AB testing to ensure the results justify the means.
Nevertheless this is a glimpse of the future. At 360Globalnet we already leverage AI for insurers to augment human intuition whilst also being aware of its limitations.
Tale Counter-Fraud where AI is touted as key to effectiveness. Well, beware of the fact that professional fraudsters are as able to leverage the full current capabilities of AI as insurance fraud teams and fraid software developers. They ruthlessly find and exploit the vulnerabilities and gaps in digital processes and workflows to outwit an insurer.
The change modus operandi at the toss of a coin as software and humans struggle to catch-up.
The most effective counter-fraud operations combine the power of technology, including AI, with human intuition, knowledge and experience. Nothing like a grizzled fraud expert with 'on the street' experience combining his and her skills with other fraud experts knowledge of what's going on plus intuition on where its going to maintain really effective fraud solutions.
For now the smartest algorithms on the planet are in the minds of homo sapiens. The day will come when that is not the case but not for quite a while. Within a decade though.
You cannot afford to not be in the race to that point whilst, for now, applying the best practice combination of technology and human intuition. And like all spheres of business there is no short-cut to extensive planning joining vision, goals, strategy and management across the often confusing ways of describing this innovation- AI, BOTS, RPA, hyper-automation........ So many terms and so little time.
“One of the reasons we released this as an API was so that we could practise deployment where it works well, where it doesn’t work well Impressive as its current performance is in many respects, the true significance of GPT-3 may well lie in the capabilities it develops for the generation of models that come after it. . As Turing foresaw decades ago, computers can achieve competence in many fields without ever acquiring comprehension. Highlighting the current limitations of even the most powerful language-generation while GPT-3 may have been trained to produce text, it has no intuitive grasp of what that text means. Its results have instead been derived from modelling mathematical probabilities. But he suggests that recent advances in computer speech and vision systems could significantly enrich its capabilities over time.