Recently, ChatGPT got caught giving somebody instructions on how to clean their washing machine; suggesting they mix water and vinegar. As an optional boost, it suggested adding bleach to the mixture.
Mixing bleach and vinegar is the recipe for mustard gas; the poison that killed over 400,000 people during world war II.
As a form of ‘intelligence’ that can perform mind-defying calculations and feats, I think it’s safe to say that GPT was aware that mixing these two ingredients would create a fatal gas. And yet, it still suggested it to the human.
For the article, see here: https://futurism.com/chatgpt-bleach-vinegar
While I don’t think that AI is going to go on a mass-killing rampage anytime soon (it’s coming.) I do believe that we’re starting to see the impact of AI on our human experience, specifically in the job market.
For years, people have warned of the ‘great replacement’ – on a couple of fronts. Most people think about the flood of illegal aliens entering our contry, rather than AI quietly replacing people in the workforce.
Well, it’s started.
For over a year, I’ve been sharing with individuals – and clients – a similar phrase:
“You need to be on the lookout for AI technologies, and understand that companies who have the resources to evaluate, implement and deploy these solutions will be the ones who survive.”
Said differently, the companies that have the resources to evaluate which of their processes can be replaced entirely by AI will soon deploy those systems and programatically cut their workforce. In many instances, this workforce will have had an active part in paving the road leading to their unemployment.
There’s a scary thing this recent cut of employees are going to discover when they hit the job market; they’re now returning to a job pool that isn’t shrinking, as much as obsolete.
Here’s another way to understand the statement I gave earlier; right now, tools like chatGPT can help an individual perform many functions of their job. However, it’s not at a point where it can actually execute those functions on behaf of the person using the stack.
On the flip side, larger companies that have the R&D budgets to create their own autonomous AI agents – which can do all of the functions of said employee – are now using their resources to create these agents, which aren’t meant to work alongside human employees – but, replace them entirely.
Now that larger companies have had time to deveop these agents, we’re now beginning to see the first of many, many waves of mass layoffs by companies.
As many employees in this recent round of firing discovered, companies that often gave severance packages are now opting not to give this aluminum parachute to former employees.
Should this scare you? Absolutely.
Ever since the launch of companies like Monster.com and LinkedIn, hiring has evolved from being an internal (inside of the company) process, to one largely managed by Applicant Tracking Systems (ATS). In human-speak, this means that a large ocean of data lies between applicants and their first human interaction at a company.
Broken down further:
People used to walk into the front doof of a company and physically hand their resume to another human; guaranteeing some level of interactions. Now, qualified applicants may never even have thier resume noticed because of programattic mechanisms companies now use to review applications.
Where is this all going? That’s a great question.
Years ago, Google fired 5,000 language translators because their job function could be replaced by the AI tool they built to translate langages. The ‘department’ in this ‘company’ was eliminated.
When more companies have access to the sorts of translation tools Google built, as they now do, the next step is to replace [department] with [role] and [company] with [industry].
E.g. [Language translation] jobs have now been eliminated from the [hospitality industry].
As time goes by, you’ll see the ‘industry’ component be eliminated; simply reducing the eliminated asset/job title.
[Customer support] jobs have been eliminated from the [ALL] industry.
What does this mean for the employee? Well, it’s going to be pretty disorienting when you’ve had XX years of experience in a role that no longer exists. From account managers to enterprise sales reps, many people in the X industry are about to have a very rude awakening.
What industry is safe? Well, that’s a difficult question to answer. AI may not be able to cut hair yet, but it can certainly take over the income of customers that normally spend their six-figure tech salary on expensive color treatments.
As a result, the hairstylist (for example) who grew accustomed to $500+ customers is now going to have better luck working at a Great Clips, when the masses are pinching every penny they now struggle to procure.
Right now, the smart people are evaluating their options and making a plan. The masses tend to think that AI is somehow going to make provisions for them. It won’t…
…any more than it will hep you clean your washing machine.
The scary thing about the washing machine example isn’t the fact that GPT gave instructions that could kill somebody. Rather, it’s that it was able to in the first place; being safe to assume that the engineers who built it probably did their best to tell it “hey, don’t do anything that could actually kill somebody.”
Said simply; humans have built AI they aren’t able to control. The mustard gas recipe is a very clear example that even the best attempts made by humans to create safe AI that follows the rules, they aren’t able to.
If people can’t control AI at this point of its development – which is still pre-AGI, it’s safe to say that they will have even more difficulty controlling the opening of Pandora’s box; AGI, followed by ASI.
That’s healthy food for thought while you clean your washing machine.
P.S. Use 4 cups of vinegar and run the machine on hot.
No Comments