Tone deaf - No problem, pt 6
It's interesting revisiting this article from 3 years ago, especially now AI products are more embedded in market.
If anything my view on AI is more polarised, bringing many new opportunites, brand new problems, as well as the existential angst many are talking about.
For me, codifying my experience with AI tools means I can help more people, while freeing up my time to do more human work. It's certainly helped with job seeker support, and increasingly so in my recruitment work.
Mainly it's giving me more time to talk to people, and when homogenous AI applications at scale are one new problem, as well as the increasing threat of deep fake interviews, in person conversations might be a necessary exnovation (just googled that - what a great word that I will now abuse at scale).
While the article below has dated somewhat, I think the threat it assesses has never been more pertinent. I've said for some time that AI might be the death of knowledge and judgement, even so far as 'meaning', something that this recent research also indicates.
Importantly the research doesn't say the recommendations were wrong, just how they were followed. Read that article and think about how AI recommendations on applications relate. I've certainly found them more problematic, so far, than the old school ATS style ranking and filtering - because of the false confidence they convey, especially when dealing with non-traditional applications. Things may improve, or so is promised.
And while words might be the first facility of LLM, they are still the core facility, given all it does is determine the most probable answer in its outputs (I know I'm being reductive, some of the developments from AI are amazing). Whereas words should be the consequence of thoughts, lead to the development of thought, and distil ideas in a way you can act on. Much like job adverts should be the consequence of effective job analysis, rather than a polished word that lacks meaning.
The greatest risk for me in recruitment isn't so much the answers AI offers, but that we become so reliant we forget to ask better questions.
I've edited down the previous article, but if you're curious, you can see it here.
September 2023
I’d commented about my hearing on a social media post on deafness.
‘Have you looked into these hearing aids - they’re designed to overcome head shadow’, came the impromptu reply from a hearing aid specialist.
Head shadow?
Not a phrase I’d heard before, despite having no hearing on one side, which effectively creates a sound-free shadow to my right.
See, when I had mumps at 18 months it killed the vestibulocochlear nerve in my right ear and damaged the one on the left.
But it was my balance that showed the nursery assistants that something was wrong - I kept falling over for no reason.
I expect they thought my slow response was normal naughty-boyness.
Balance, intermittently poor audio processing, energy required to maintain attention in noisy environments and that bloody tinnitus - all these issues are far more annoying than not being able to hear in one ear.
That and not being able to tell where sounds come from.
“Over here, Greg!” is the most useless call on a football pitch.
And the head shadow thing? Well, it’s never been an issue - I just turn my head when in conversation. It might look like I’m turning to kiss you, but it’s just a conversation I want honest.
Sure I miss out on certain audio cues, but it’s rarely an issue.
Anyway, curious as I am I looked into the £4,000 hearing aid that solves head shadow by placing a microphone on the deaf ear and projecting it into the the good ear.
Quite the brute force tech solution and one I’m sure some benefit from.
But the key issue is one of reverb. While the product is good, it would impact the quality of sound in my good ear, which seems a pointless trade-off for me, when my process of life is more effective than shiny tech.
It’s ironic that I consider myself highly excitable when it comes to technology developments, yet I’m never an early adopter.
Because I always ask myself, other than giddiness, what problems does this solve and how will it enhance my life.
And so it is for AI in recruitment, and the shininess of tech in general.
They are positioned as solving problems, and indeed they do, but in many cases, it’s the wrong problem that detracts us from the real opportunities for better processes.
Simpler administration, speed and scale are clear benefits.
And some of these AI-branded automations are excellent.
The Star Trek-style auto-translator makes me giddy. As does the sales bot conversation-as-a-human sales video that’s doing the rounds.
But AI for writing? What problem does that solve?
For me, it depends on what you consider sales to be.
Is sales pitching a generic message and hoping someone will bite?
Or is sales laying out the facts so that the right person can reach an objective decision? A decision which might mean getting in touch, or changing their view on what needs recruiting.
For the former AI writing seems to solve a problem. The words are certainly better, but the message is even worse.
Yes: prompting. But if you can prompt the right message, why aren’t you just writing it, so that you embed that knowledge for use across other mediums - conversations, meetings and so on?
While AI writing does not meet the requirements for the second type of sales. You can only do this by establishing the right facts from a briefing, consultation or interview. Then interpret this with meaning to the right people so that it creates the right impact. Sales as data visualisation, if you will.
And if AI doesn’t work for sales in this way, why would it work for content or marketing?
I think the answer depends on whether you value contextual insight or not. Until AI is better at asking questions that create meaningful conversation, it’s a non-starter for me.
I say until because this hesitation will be moot in the next year or so. Because it isn’t what AI is now that’s the opportunity - it’s what’s coming soon. I can’t wait to see what’s next.
Until then, AI, for writing, is a question of in one ear and out of the other. Which, if you’re deaf in the one, is quite the problem.
Go learn copywriting instead, and listen to how it transforms your approach to recruitment. Then layer tech on top, when it’s ready.
Thanks for reading.
Regards,
Greg
p.s. If you like good recruitment and you need help with it in the UK, drop me a line. DM on LinkedIn, email greg.wyatt@bwrecruitment.co.uk, whichever you prefer.

