Attack Of The AI
5 in 5 - Brave & Heart HeartBeat #128 ❤️
This week we’re doing a deep dive on AI in its various forms and uses. From pure tech and AI for the sake of AI, to real-life utilisations from recruitment to gambling.
Is it actually useful and worth developing, is it a tech vanity project, and will it inevitably always end in some terrifying nonsense like that dance scene in Ex Machina?
Let’s get into it.
Were you forwarded this? Not a subscriber? 👉 Sign up here
#1 - What (Or Who?) Is Google’s New AI Chat?
Lamda is its name – Language Model for Dialogue Applications, and it is an AI technology used to power chatbots which has just been launched in the UK for testing.
Google has opened up the app with a limited trial to allow users to interact with it, however, to avoid any sweary moments the AI will NOT be able to learn from users, unlike others who have been burned in the past…
This may be ringing a bell for you, as it’s the same AI which a former google employer, who was promptly sacked, claimed was sentient.
BBC tested it out with the only three available scenarios in the trial – asking the Lamda to use its imagination, asking it to give a to-do list and break down the actions needed to complete tasks, or, weirdly, to talk about dogs.
They described the response as being more sophisticated than any other AI chat bots out there, and weirdly verbose, like talking to Stephen Fry.
While Google STRONGLY deny that Lamda is sentient, despite Lamda literally saying that they were (controversial) they have yet to refute any claims that they have stolen the consciousness of Stephen Fry, so watch this space.
#2 - What Is Up With Elon Musk’s Robot?
Elon Musk, at once global laughing stock and tech genius, has had a bit of a flop with this one.
He’s been promising a human robot called Optimus for over a year, however the two prototypes he recently unveiled didn’t exactly razzle dazzle us. The most advanced robot, gave an unsteady wave and was then pushed across the stage by three actual people.
Musk claims that the eventual release of “Optimus” will mean a “future of abundance, a future where there is no poverty, where you can have whatever you want”. Describing it as “a fundamental transformation of civilization”. Apart from the fact that he recently watched iRobot, what does that actually mean?
Melonee Wise, CEO of Fetch Robotics, who make robots which operate in warehouses alongside human workers says - “The idea that this will be doing anything useful in five years is laughable”, noting that the claim that Tesla was using Autopilot, its driving assistance system, to control the robot was “perplexing” due to how different walking is from steering a vehicle.
To the casual observer, your average Joe Not-A-Robot-Expert, it also seems pretty much on the same level, if not worse, than the robot Honda showed us literally ten years ago.
Watch this space, for a good laugh if nothing else.
#3 - Midjourney – Greatest Artist Of Our Time
This year, like every year, the Colorado State Fair held their annual art competition. However this year, the winner of the digital art category was a Mr Jason Allen, who barely lifted a finger to created his artwork. He made it with Midjourney, an AI program that turns text into hyper-realistic graphics.
This created a huge backlash as graphic designers everywhere accused him, shockingly, of cheating.
A.I.-generated art has been around for years, however certain tools released this year — i.e. DALL-E 2, Midjourney and Stable Diffusion — have made it possible for pretty much anyone to create complex, abstract or photorealistic works by typing a few words into a text box.
You provide the text – literally seconds later the tools provide the image.
There’s debate online about how worthwhile this art actually is, with some arguing that it still takes human creativity to come up with the text prompts, and others saying that new tech in art is always criticised, even the camera was viewed with horror by some artists when it was first invented.
Others are worried that the actual career of graphic designer is going to be stolen by AI tools, with Getty Images literally banning the sale of AI generated art.
What’s the verdict? Personally, we think Mr Allen cheated, and should give his $300 prize money to anyone capable of creating anything resembling art on Microsoft Paint – because that is real talent. (We vote for this guy).
#4 - AI Recruitment Bias
While a 2020 study says that one quarter of the 500 companies questioned were using AI for talent acquisition in the form of automation, analysing candidate videos and applications, the experts have deemed that this use of AI is nothing but “pseudoscience” – and that is not the kind of science you want to be giving a hand in your recruitment.
The use of AI to analyse a candidate’s speech and bodily movements in recruitment videos or to “see through to your personality” just by looking at your face has been touted as the modern version of “phrenology” – the false (and super-duper problematic to boot) Victorian theory that skull shape could reveal character and mental faculties.
While an expert says that AI can help increase diversity by filtering from a larger candidate pool, it can also miss out on candidates if the rules and training data are incomplete or inaccurate. As seen in 2018 with Amazon, who scrapped the development of an AI-powered recruitment engine because it could detect gender from CVs and discriminated against female applicants.
With the software being in its infancy, there are apparently both opportunities and risks, but we reckon this one could probably be left well alone for the moment, while recruiters should focus on the basics – hiring to values, real connection with the recruitment process, and not getting distracted by nonsense tech.
#5 - Australia Rolls Out Facial Recognition To “Help” Gamblers
According to a recent announcement, pubs and clubs across Australia will be installing cameras that can capture biometric data of everyone who enter.
The point of this is to crosscheck their face with a database of “problem gamblers” who have signed up – basically a list of “barred” gambling customers who want to be banned from their own accord. In the event of a match, a staff member is notified to that they can “intervene” and refer the person to support services.
According to the executives behind the rollout, 85 percent of problem gamblers support the move – although this survey is nowhere to be found. Experts globally however have long warned against the use of facial recognition technology in this way.
Among them is Samantha Floreani, a program lead at Digital Rights Watch Australia, judges that facial surveillance tech is “invasive, dangerous”, and that “we should be exceptionally wary” of introducing it as a “quick-fix” harm reduction strategy.
What do we think? As we don’t 100% know where the data is going or what it’s being used for, the privacy issues behind it are pretty concerning, especially if we consider those with gambling addictions to be at-risk members of the public.
Brave & Heart over and out.
Bonus
Photographer Creates AI Girlfriend to Placate Nosy Relatives
Are your family always on your back about finding someone, getting married, settling down, but you’re just out there living your life?
This photographer’s remedy to nosy aunties was the creation of an AI generated girlfriend to pose with him in photos and stave them off for a while.
We can’t build a better robot than Hyundai, or paint better than an art bot, but we can help you recruit smarter without needing to use AI.