How GOOD or BAD is AI?

[AviatorTom]

Rishi Kant, *Reuters* (01/29/25), via ACM TechNews, reports that...

"An audit by trustworthiness rating service NewsGuard found the chatbot
rolled out by Chinese AI startup DeepSeek had an accuracy rate of 17% when
it comes to delivering news and information. DeepSeek provided vague or
useless answers 53% of the time and repeated false claims 30% of the time,
with a fail rate of 83%. In comparison, its Western rivals, including
OpenAI, had a 62% average fail rate."

So, how GOOD, or maybe we should ask how BAD, is AI?

[SomothErgai]

According to an AI

AI, like any tool, varies in quality depending on the model, its training data, and how it's deployed. The question of whether AI is "good" or "bad" depends on context.

1. Accuracy and Reliability:

AI is not universally unreliable. The NewsGuard report specifically highlighted DeepSeek's chatbot as highly inaccurate (17% accuracy), but other AI models, like OpenAI's ChatGPT and Google's Gemini, have significantly higher accuracy rates (averaging around 62% fail rate, meaning 38% accuracy).

Studies, like those in medical research, show that well-trained AI models can provide highly reliable information when properly developed and tested.

2. AI’s Strengths:

AI is incredibly efficient at processing and analyzing vast amounts of data quickly.

It automates tasks in journalism, research, coding, and other fields, improving productivity.

AI-powered tools assist in medicine, cybersecurity, and fraud detection with high success rates.

3. AI’s Weaknesses:

AI is only as good as its training data and safeguards. Models like DeepSeek struggle with misinformation because they may be biased or insufficiently filtered.

AI lacks human intuition, meaning it can misinterpret context, leading to errors in critical areas like news reporting.

So, how good or bad is AI?

AI is neither inherently "good" nor "bad"—it depends on how it is trained, used, and monitored. Poorly designed AI (like DeepSeek, based on this report) can spread misinformation, while well-developed AI (like medical diagnostic tools or research assistants) can provide highly accurate and useful insights. AI isn't a monolithic entity; its reliability depends on The model and its application.

[stupefy]

50/50 as in 50% wrong to 50% right is the worst you can do with a random coin toss or flip of a boolean value. That is considered a failing grade in most computer science courses. For example, in the course I was enrolled in, if your software could not at least achieve 75% correct response, you were not even allowed to submit your work for review. Typically you haven't done a good job until you hit at least 99% right to 1% wrong. When you get to 99.7% right to 0.3% wrong you've done a really good job. When you get to 99.9% right you've done an exceptional job. On my first AI project in college I used bayesian inference to achieve 99.7% correct prediction of a particular disease state from medical diagnostic information and 99.9% right conclusions to predict a pre-disease state that would eventually lead to the disease if habits were not changed to improve health in particular areas. I wrote the software in about 30 minutes. So with a person that has good intention and skill, you can achieve good results. Anything worse than 50/50 is considered intentionally bad. It means the author of the code has intentionally trained their software to mislead people. That is what many large AI projects are actually designed to do. They want to actively mislead people. They want control over their minds. AI is being used as a legacy media alternative to brainwash people over the internet. Billions of dollars are being spent to engineer bad software that lies to you.

[stupefy]

I don't know why more people don't understand this basic principle. If someone asks you a yes or no question and you're not even trying, you will get the answer correct half of the time. So if they get it wrong more often then that they are intentionally being malicious and trying to deceive you. The same goes for AI. People understand that there are computer viruses, Trojan horses, worms, key loggers, fork bombs, etc. and they go out of their way to buy antivirus/anti-malware software to protect them from all the bad programs people write to harm them, but then AI comes along and nobody questions it. Bunch of fools.

[purplepoppy]

I don't trust AI. The main gripe I've got with it is if I look up a topic I know a lot about on Google I'm seeing more and more written by AI and 99.9% of the time the information is nonsense or so vague as to be pointless. So why should I trust information on topics I don't know anything about? And that's is the danger, too many people are blindly trusting whatever it churns out.

[sage2021]

I just learned yesterday that AI was able to find the exact medicine for a man who was dying of a rare disease.
Out of hundreds of medications, AI was able to find the one that he needed to cure his rare disease and now he's alive and doing well thanks to AI🙂

[Likes2drive]

I doubt it’s good and probably not a good idea using it

[OneViewpoint]

Given that we live in a world where half the people are more stupid than average, how dangerous is it to tout a machine as "intelligent". How much will stupid lazy people rely on that without questioning or researching the output?

The truth is that Artificial Intelligence is a label, branding, not a factual statement. AI is not "intelligent". The algorithms can mimic certain aspects of intelligence to a certain degree. But the algorithms can also be manipulated. Gee, who would do that and why?

At the point the stupid half of the population is being manipulated, doesn't know and doesn't care, does reality just become whatever subjective thing whoever controls the algorithms wants it to be? Is there any danger in that?

Let's say there's no evil conspiracy and AI doesn't become an out of control Sky Net from the Terminator movies. But numb nuts rely on an AI's advice for investing and lose 95% of their 401K because the AI lied but doesn't have any concept of lying or good and bad. Or maybe an AI was used to rip off someone's, proprietary image, like Brad Pitt. Is there any legal recourse? Against who? The employee who wrote the algorithm? The company who owns the AI program like Chat GPT, Open AI, Microsoft, Google? Investors? Do you sue the AI itself? Is AI awarded "Personhood" like our demented courts gave to corporations.

There are currently zero laws on the books to define how a lawsuit involving AI would be resolved in our court system. Our government is once again 20 years behind our technology development. And our legislature is now filled with unqualified morons who are barely literate with no legal background. So, the lawyers hired by corporate lobby groups will once again write the laws.

Does any of that sound dangerous?

[ManOnFire]

There are both good and bad aspects to it, although I lean more towards bad than good. AI can imitate people, create images and videos that look 97% real (they've gotten a lot better at it now), and can have their own voices. To me this is largely bad as it opens doors for both deception and the possibility of wrongly incriminating someone in the future.

On a personal note for me, there is something oddly satanic about many AI videos and reels out there now, and something about them can even feel somewhat nauseating to me.

[FreyaRed]

Maybe one is better programmed and learned than another.
However I think in current status of AI estimating AI's "intelligence" that relies 100% on quality of external sources isn't more than marketing gibberish.
What is interesting Chinese AI can run locally on a newer workstation while OpenAI needs supercomputer hardware with several professional Nvidia GPUs. This make a lot of more computing operations per second for OpenAI, but is this model that much better? I don't think so.

[EzraIsShy76028]

I mean in THEORY AI is a good thing because like most technologies on paper it's designed to make our lives easier, faster, and more efficient. But in ACTUAL practice it seems like a recipe for disaster as AI is in theory a perfect tool being used by imperfect humans. Lots and lots of people are gonna lose their jobs due to AI, for sure. And they already are. And art and music will even further be ruined. Not to mention deepfakes and shit that could be used against people. So in the long run it probably would be better if AI didn't exist. Or at least not any more advanced than simple AI chat room bots for fun. But nothing bigger scale that'd steal people's jobs or anything.

[AnonAndrew]

AI is overall pretty bad.

It's good for generating random text and images, and can be used to help write things after a standardized format.

It's also good at many calculations, and at finding basic information.

However, anything beyond that, and I find it messes up things all the time.

I tried using it for specialized areas of my work, but I had to give up pretty quickly. It was as if the AI were a stubborn kid who knew nothing, but didn't want to admit it, so it just made up random answers that weren't true.

[JustinTheGreat]

There’s some good uses for AI but a lot of it’s bad.

AI is bad for the environment. I couldn’t understand how at first but apparently generating a single AI image takes as much power as what’s needed to charge your phone to full battery. So it releases a lot of emissions.

AI steals jobs. Why should companies hire artists when a computer can generate their ads for free? Even Coca Colas iconic Christmas commercial was AI generated last year. And to create AI images, they use artwork from actual art they find online. So this AI artwork is essentially being stolen from actual artists.

AI is dangerous when it comes to misinformation. You can create an image or an audio of someone doing or saying anything. It’s becoming harder and harder to tell if something is real or AI. Of course now they are creating programs that can detect it. But a lot of people can’t tell the difference, especially the older generation.

[Anonymous]

AI in itself is a double edged sword.

There are amazing ways in which AI can be used, such as finding a cure for life-threatening diseases (AI can't actually use the specific cure though) or helping students finish their coursework.

However, it can be used very negatively as well, such as replacing people in jobs (this is more evident in fields such as software, technology, or computer science than in fields such as construction).

Sadly, we see AI being used for negative impact far more often than we do for positive impact, at least in the case of the powers that be.

[marfaret]

it's benefits are more than it's cons actually, seems unbelievable they have they own liberty and authority, by the way they made by humans so they can't cross the line. also works are pretty easier with them and people can more unwind within their jobs, but it can't be prevented that AI got many options of our youth to present them self in society.

[jshm2]

It's neither good or bad. Like any "tool" there are many factors to take into consideration. The quality, the cost, the support, etc etc.

Right now, there is no 100% quality AI. The technology is relatively new and everyone is trying to find ways to move it forward.

[jahaims]

It's neither. It's just like any technology. There will be pros and cons. A vehicle is very convenient and helps the economy because the average person can travel for work and trucks can transport goods. Car accidents and pollution would be the cons of having vehicles.

[dragoblack]

In terms of replacing jobs and over complicating basic needed actions it is absolutely terrible and a waste of resources & energy. In terms of it being used to assist basic functions, it's perfect. AI doesn't need to be powerful or insanely smart to be good.

[jennifer_bloom]

I believe this claim about China might be exaggerated if not totally made up since American businesses might lose billions if China steals their AI consumers. I have not experienced AI in a thorough way but being Chinese in Canada and seeing previous lies about my ethnic group by others, I am suspicious of all collaborative claims made about Chinese businesses by competing groups.

[OlderAndWiser]

There may be some limited circumstances in which AI would be useful. G@G is NOT one of those circumstances!

[RHK36]

It's like the Internet itself. It'll be as good as we use it. The same as anything else we created technology wise that was a major game changer. How people in society use it will determine if it was good or bad. Check back in 10-20 years

[NatalieKeller95]

it's horrible and should have been stopped before it got to this point, it's at the point of no return now, I fear what the future might bright with AI growing more power

[naartjie]

Can we all just agree we should destroy everything and live like they did in the 70s? Please and thank you.

[beefcakebradybatson]

He's in the Basketball Hall Of Fame, though his great career ended sourly with the Pistons.(Wait, you mean Artificial Intelligence, NOT Allen Iverson!!! AI can be a boon or a menace to humanity depending if it's used for good or evil.)

[lorenzomichael]

Well, it all depends if I could teach AI to wash my car straighten out my garage and do my laundry. Then I think they’re awesome. If not that then I think there worth the shit.

[Anonymous]

Neither. AI is a useful tool but we need to be aware of its limitations.

[Joshydavid25]

I guess we don't have to worry about Skynet any time soon..

[exitseven]

AI will evolve into something that will ultimately be an important tool for mankind.

[bolverk]

Any AI is like any program, it is only as good as the information used in its construction, and the use to which it is put.

[genericname85]

like so many tools in the world, it's good or bad depending on who uses it and for what.

[worldscolide]

The moment AI becomes sentient, we have problems.. Until then it's just a tool..

[Crazzyyworld]

right now they are bad because they take our job but then it will be better when they start working for us leaving everything for free.

[Staximus]

Stephen Hawking said AI is the beginning of the end for our species

[Ethan_T]

I think AI is cool but should probly be just for fun

[Enginn]

Its good when its used for good purposes. Bad when its a bad purpose. Technology makes life easier so that should be the main focus.

[Sabretooth]

ai is like a child--it'll often be what you make of it.

[Sabretooth]

We can't let become like the Terminator franchise. That's a loser for both sides of the war.

[Anonymous]

I've made about a 1000% off my Nvda stock over the past couple years. AI is VERY GOOD!

[SimpleGirl07]

it’s rise of skynet.

[Magda1992]

It’s Starting?

[grega239]

Its both

[Bklynbadboy12]

I hate it

[ienjoylookingatgreen]

it's pretty good