8 Pros and Cons of The Rise of AI
When this story was dropped in my lap I initially couldn’t find any real problems with Artificial Insemination apart from perhaps some sort of eugenics program. I still managed to come up with 5000 words on the subject. After my editor saw it and informed me he meant the other AI the job got a lot easier. With the launch of GPT3 (shortly superseded by larger incremental numbers) came the hyperbole machine. AI is here, and it’s going to take everything from you. Your job, your wife, your civilization, possibly even your burgeoning art career. But on the flipside it could probably make your life a lot easier for writing the most beige Best Man speech ever conceived in mere seconds! So it all balances out really.
All joking aside, let’s take a look at some of the pro’s and con’s of AI, and decide once and for all whether opening this pandora’s box was a good idea or not. As if we have some sort of say in it.
1. Medicine Just Got Easier
Perhaps the greatest thing AI and machine learning can do is improve health outcomes. What’s better than happier healthier lives? Honest question! I’m having trouble with those two and if there’s any fallback options that’d be great.
Machine learning can quickly test and analyse hundreds of samples, and do end-to-end drug discovery far quicker than humans ever could. Developing new drugs and treatments can take a decade and billions of dollars. But already with the aid of AI a couple dozen treatments are going through clinical trials already. Startups are cropping up all over the world, and scaling quickly. It would be an exciting time to be an investor in that space.
In practical terms AI can do contact tracing for future outbreaks. All it needs is the barcode on the back of your head.
Patient care could be streamlined, aiding GP’s with cross referencing similar cases and symptoms to yours and finding the most likely successful outcome. In worst case scenarios AI could aid in remote care when a human isn’t available.
2. War Machines
Prince Harry once compared firing hellfire rockets from an apache to playing games. “It’s a joy for me because I’m one of those people who loves playing PlayStation and Xbox, so with my thumbs I like to think I’m probably quite useful,” upsetting the Taliban in the process. Soon even gamers won’t be participating in wars. It’ll be a bunch of logistics guys in a warehouse somewhere releasing Boston Dynamics dogs with guns strapped to their backs. Stuart Russell, Professor of Computer Science and founder of the Center for Human-Compatible Artificial Intelligence at the University of California, has described possible scenarios including small drones that could wipe out any potential target from a city automatically. For instance “Anti-personnel mines could wipe out all the males in a city between 16 and 60 or all the Jewish citizens in Israel and unlike nuclear weapons, it would leave the city infrastructure.”
Well, I can at least take great comfort knowing that the world I leave behind after a drone gets me won’t be just molten glass.
3. Data Control
On the flipside of AI aiding in health is that for this to be optimal health institutes would be required to work on the same systems. Data sharing would be imperative for proper AI analysis. There are massive pushes for this very thing currently. Patient data would be available everywhere, so you’d have to weigh up how comfortable you are with that, and what it would mean when this data is inevitably breached. If the data attached to you could help save a life in some small way though how can you protest? Add the contact tracing AI and already existing trackers baked into your google account and android and AI models are starting to get a very clear picture of who you are and where you are. Machine learning could also extrapolate where you WILL be. Let’s face it, that’s not a hard one. You’re either going to work, at work, or coming home from work. If you lived in a country that wasn’t particularly amenable to who you are based on your race, sexuality, or religion then it could become even more problematic when they start buying up those anti-personnel drones.
4. We Can’t Trust What We Read Anymore
Do you remember a time when everything on the internet was impartial, infallible, and most importantly true. If you do, I’ll have whatever you’re smoking. It sounds like a relaxing time. AI that takes advantage of large language models can generate just about anything you like quickly and efficiently. It might not be right or objective, but it’ll very confidently sound like it is. Setting up a legitimate looking news site with this tech would be simple and easily reproduced, creating a lot of white noise online. Bending the angle of news stories based on personal bias or affiliations would be just as easy as… It is now I guess.
Just as likely is that users using the content from these generators verbatim are just as vulnerable. GPT-3 has already shown in studies to be a left leaning libertarian, and less restricted models happily generate racist speeches. A complete lack of understanding on what GPT-3 even does has even led to headlines about a lawyer using GPT-3 to bang up his court filings. The generator spat out a bunch of legalese which sounded good, but cited a number of fake cases and fake quotes. This completely torpedoed his case, and the court’s time.
As record numbers of kids are now shilling GPT-3 output as answers to homework, how much of its generated opinions will become held beliefs of the impressionable? How much does Pepsi need to pay OpenAI to make the model always say that Pepsi is better than Coke?
5. Finding The Aliens
Humans may not be the ones to make first contact with aliens, it might just be an algorithm we set to the task. Current searches for Aliens undertaken by SETI groups primarily focus on scanning the skies for what’s known as technosignatures, non-naturally occurring signals coming from space. If there’s one thing machine learning is any good at, it’s pattern recognition. Conversely, can you think of anything more terrifying than a robot on the other side of the milky way pinpointing earth via old Radio Hauraki transmissions. (Most of our radio waves would be undetectable once they’ve reached the next closest star just 4.4 light years away.)
6. Boosting Crime
Everything is easier with a massive artificial intelligence on your side, and that includes crime! Phish attacks or other social engineering hacks online could be done at a greater scale if everybody is getting an individually crafted response leading them to believe that “yes, this Nigerian prince is the real deal.”
Impersonating people has also gotten incredibly easy with advances in both deepfaking faces and voice manipulation software. All a criminal needs is some audio of you online to then fake your voice, and then they’d be able to trick your mum over the phone to wire them money.
7. Economic Growth
For once Tech isn’t running trades out of town. If you’re a plumber you’ll be fine. If you sit on your bum all day in front of a PC, chances are it’s coming for you. In 2019 Joe Biden told coal miners facing job insecurity to “learn to code”. Hopefully nobody wasted their time on that one because I got Chat GPT to write me a program in about 10 minutes. It looks like coders might need to learn how to mine. General statistics show that around 3-14% of the world’s workforce will have to find a new gig. But a Goldman Sachs report believes that this should shake out to a 7% boost in GDP so I’m chalking this one up to a pro! Sorry about your job loss but Zuckerberg’s next boat isn’t going to buy itself.
But in seriousness, we’ve gotten pretty adept at doing new things. Based on job descriptions 60% of jobs in the States did not exist in 1940. There are a few things AI can do to a job, automate it, or augment it. Augmentation would mean that it works as a tool to make your job easier, while automation would mean that your employment as a data entry guy is goneburger. Finally it could also create new jobs in human oversight and AI wrangling.
8. Creative Dilution
Somewhat myopic in the grand scheme of things but anyone who has a very idiosyncratic creative style is now easy pickings for AI generators. I can’t scroll two feet without bumping into some disgusting Wes Andersonesque abomination. AI generated art is like photographs of me visiting the Eiffel Tower and pretending to hold the top of it, creatively bankrupt.
Vaguely popular Illustrators, notoriously underpaid as it is, are finding the internet flooded with fake knock offs of their work. To recreate somebody’s style in something like Stable Diffusion only takes 20 or so high quality images for it to get to work exporting an infinite amount of iterations of their stuff. For personal hobbyists this is a bit of fun. But artists are finding google search results crammed with the knock offs, messed up hands and all. The quality control is terrible, and means that people finding their work for the first time could be turned off. If even one client decides that “I’m not hiring a guy that draws their characters with fingers that look like an octopus.” then it’s already done some damage.
AI-generated books with zero quality control are beginning to clog competitions and Amazon books. Cutting through the noise as a creative in any field just got that little bit harder.
Pros of AI According To Chat GPT
Automation and Efficiency
AI can automate repetitive tasks, leading to increased efficiency and productivity. This can free up human workers to focus on more creative and complex aspects of their jobs.
Improved Decision-Making
AI systems can analyse vast amounts of data and make data-driven decisions more accurately and quickly than humans. This can lead to better outcomes in areas such as healthcare, finance, and business.
Enhanced Safety and Security
AI-powered systems can be used to enhance safety and security measures, such as detecting and preventing cyber threats, monitoring public spaces for suspicious activities, and predicting natural disasters.
Advancements in Science and Research
AI can assist in scientific research by processing and analysing large datasets, identifying patterns, and discovering new insights. It can accelerate progress in fields like medicine, genetics, climate science, and space exploration.
Personalised Experiences
AI enables personalised experiences by analysing user data and preferences. This can be seen in recommendation systems, personalised marketing, and virtual assistants, which tailor content and services to individual users, enhancing user satisfaction.
Cons of AI According To Chat GPT
Job Displacement
As AI automates tasks previously performed by humans, there is a risk of job displacement. Workers in certain industries may find their skills outdated or unnecessary, leading to unemployment and income inequality.
Ethical Concerns
AI raises ethical concerns regarding privacy, bias, and accountability. Issues such as the misuse of personal data, biased algorithms, and lack of transparency in decision-making algorithms need to be addressed to ensure responsible and fair use of AI.
Dependence and Reliability
Relying heavily on AI systems can make society vulnerable to system failures, errors, or cyber-attacks. Overdependence on AI without sufficient backup plans or human oversight can have serious consequences.
Lack of Human Touch
AI lacks human emotions, empathy, and intuition. In fields like healthcare or customer service, the absence of human interaction and understanding can lead to dissatisfaction or the inability to address complex emotional needs.
Potential Security Risks
AI can be exploited by malicious actors for nefarious purposes. There is a concern that AI technologies, such as autonomous weapons or deepfake techniques, can be misused, leading to security threats and erosion of trust.