Above: The panelists at the Chamber of Commerce discussion on Artificial Intelligence in HR.

BitDepth#1408 for May 29, 2023

Some things are actually simpler than they seem. Figuring out what’s likely to happen to your job in the face of artificial intelligence (AI) advances may not be as hard as you think.

In an interview with a prominent petroleum sector executive many years ago, I was told – by someone who had negotiated multi-billion dollar deals that fundamentally changed the fortunes of this country – that what was necessary was “shopkeeper logic.”

“You sell for more than it costs you to make or buy something.”

When it comes to AI, the basic math comes close. At this point in the development of the technology, any task that does not require interpretive human intervention and is executed with repetitive non-creative input is likely to be fair game for AI automation.

According to Abigail Bynoe, Deputy Permanent Secretary in the Ministry of Public Administration, the government is optimistic about the prospects for AI in governance and is working on a legislative framework to support AI while developing a National AI strategy.

Bynoe was speaking at a Chamber of Commerce Webinar, Artificial Intelligence and H.R – Implications to the Workplace on May 18.

The Public Administration Ministry is involved in the digitalisation of personnel records and updating job roles and specifications while considering appropriate training for the public sector, so the role of AI as a replacement for some routine governance tasks is likely to affect the balance of staffing in government.

“If you program it right, it can do useful things, but the services must be aligned,” Bynoe said.

Programming generative AI software usually means feeding it large swaths of data for it to use as a resource when it is queried.

According to Rishi Maharaj, a Data Protection Advisor with Privicy, there are questions of digital human rights that arise from such use.

“[Companies must] understand the risks that are involved in making use of company generated data,” Maharaj warned, “particularly data that captures personally identifiable information.”

Darren Mohammed, Senior Manager at MicrosoftTT was even blunter.

“Don’t feed company data into a [public] large language model,” he said, noting that tools on Microsoft’s cloud platform, Azure, allow a company to containerise company data collection.

Microsoft’s recent annual report on work trends, Will AI Fix Work, was released on May 09 and identifies three trends that the company has identified in the way AI is likely to be used in the immediate future.

Microsoft believes that AI tools can relieve what it describes as the “digital debt” that’s now part of every working day, the blizzard of information, emails, meetings and notifications that obscure concentration and focused work.

“This new generation of AI will remove the drudgery of work and unleash creativity,” Satya Nadella, Chairman and CEO, Microsoft promised in the report.

To that end, Microsoft is introducing AI tools to its Office productivity suite and other products as ‘co-pilots’ for the working day.

The new tools leverage the mindshare in the company’s partnership with OpenAI, which began with the incorporation of ChatGPT in the Bing Search engine and now incorporates Microsoft’s homegrown AI tools.

In the new copilot concept, AI chat options will become part of the Office apps, the Edge browser, security tools and the OS level search functions in Windows 11.

Microsoft has its own Azure based large language models (LLM), most notably the Megatron-Turing NLG 530B, introduced in 2021 as “the world’s largest and most powerful generative language model,” a high-end research project with Nvidia to explore the scaling potential of LLMs.

It isn’t clear where AI training will take these models and how capable the resulting tools will be, but what’s already available is already disturbingly impressive.

How we think about education’s role in creating employable workers, how employers embrace these tools to either assist, elevate or replace their staff and how legislation limits or engages these tools are not questions for tomorrow, they are the front burner questions that need to be continuously engaged.

“Learning isn’t keeping up with the pace of work,” the Microsoft report notes.

How a country like Trinidad and Tobago which uses the infrastructure of government to create widespread employment engages tools designed to streamline the workforce will be questions that become relevant more quickly than we expect.

Taking charge of this rapidly evolving scenario of workplace change will demand one fundamental and overdue evolution in governance, the continuous gathering and distribution of actionable information about how this country operates.

It was a note that Jonathan Cumberbatch, Assistant VP, Human Resources and Administration at UTT touched on cautiously when he noted that, “Data drives most of the conversation outside of TT, but we don’t have a sense of that in TT.”

The propensity of governance to proceed on feelings, hunches and political expedience might have worked in the past, but the national distaste for transparently gathered, publicly available information cannot continue into an era hallmarked by a reliance on reliable, continuously updated datasets.