AI's Thoughts on AI

JimFord047

Whilst a Lot of People are worried about AI Moderating in-game Text, This game is not alone in using the technology. So for a Bit of FUN, I asked Co-Pilot Pro, a few questions that I have been asked about AI (please note, these are general answers from an AI, and not the Specific Setup of the AI used within this or other games, but it will give a little Peace of Mind) , and here are it's answers!!!


"


how accurate is AI

Sent by Copilot:
The accuracy of AI can vary significantly depending on the specific application, the quality of the data it is trained on, and the algorithms used. Generally, AI systems can achieve high levels of accuracy in tasks such as image recognition, natural language processing, and predictive analytics. For example, some AI models can identify objects in images with over 99% accuracy1.

However, AI accuracy is not always guaranteed. Factors such as biased data, insufficient training, and complex real-world conditions can affect performance. In some cases, AI systems might produce errors or biased outcomes2. It’s also important to note that while some AI models are highly accurate, they may lack transparency, making it difficult to understand how they arrive at their decisions1.

Ensuring AI accuracy involves continuous testing, validation, and improvement. Researchers and engineers often use techniques like mathematical proofs and robustness testing to enhance the reliability of AI systems2.

Do you have a specific AI application in mind that you’re curious about?

1
hbr.org
2
scienceexchange.caltech.edu
3
phrasly.ai


Sent by you:
how accurate is AI in understanding human conversations


Sent by Copilot:
AI has made significant strides in understanding human conversations, especially with the development of advanced models like GPT-4 and other transformer-based architectures. These models can comprehend context, maintain coherent multi-turn dialogues, and generate responses that are often indistinguishable from those of humans12.

However, the accuracy of AI in understanding human conversations is not perfect. While some AI systems can achieve near-human performance in specific tasks, they can still struggle with nuances, sarcasm, and context that require deep understanding and common sense3. For instance, AI models might misinterpret ambiguous statements or fail to grasp cultural references4.

Overall, AI’s ability to understand human conversations is impressive and continually improving, but it still has limitations. Continuous advancements in natural language processing and machine learning are helping to bridge these gaps2.

Is there a particular aspect of conversational AI you’re curious about?

1
link.springer.com
2
ibm.com
3
inspire-ted.com
4
technologyreview.com
5
blogs.microsoft.com
+1 more
Sent by you:
is AI used for Spying on in game text
Sent by Copilot:
Yes, AI can be used to monitor in-game text. This is often done to ensure compliance with community guidelines, detect toxic behavior, and prevent cheating. Game developers and platforms use AI to analyze chat logs and in-game messages for inappropriate content, harassment, or other violations of terms of service1.

However, the use of AI for monitoring in-game text raises privacy concerns. The ability of AI to analyze and interpret large volumes of text data can lead to more comprehensive surveillance, potentially infringing on user privacy2. It’s a balance between maintaining a safe and enjoyable gaming environment and respecting players’ privacy.

Are you concerned about privacy in gaming, or are you interested in how AI improves the gaming experience?

1
arstechnica.com
2
malwarebytes.com
3
gamedesigning.org


Sent by you:
which games use AI to monitor in-game discussions


Sent by Copilot:
Several games use AI to monitor in-game discussions to ensure a safe and enjoyable environment for players. Here are a few examples:

League of Legends: Riot Games uses AI to monitor chat for toxic behavior and harassment. The system can issue warnings, mute players, or even ban accounts based on the severity of the behavior1.
Overwatch: Blizzard employs AI to detect and address toxic behavior in chat. The system helps maintain a positive gaming experience by identifying and penalizing players who violate community guidelines2.
Fortnite: Epic Games uses AI to monitor in-game chat for inappropriate content and behavior. This helps in maintaining a safe environment for its large and diverse player base3.
Minecraft: Microsoft uses AI to monitor chat in Minecraft, especially on public servers, to detect and prevent bullying, harassment, and other forms of inappropriate behavior4.
These systems are designed to balance the need for a positive gaming environment with respect for player privacy. Are you interested in how these systems work or have any concerns about privacy in gaming?

1
builtin.com
2
gamerant.com
3
en.wikipedia.org
4
leewayhertz.com
5
media.gettyimages.com
6
media.gettyimages.com
+2 more


Sent by you:
how reliant is AI upon human parameter inputs


Sent by Copilot:
AI systems are significantly reliant on human parameter inputs, especially during the development and training phases. Here are some key areas where human input is crucial:

Data Labeling: Humans are often needed to label and annotate data, which is essential for training supervised learning models. This process helps the AI understand what it should be looking for in the data1.
Algorithm Design: Human experts design the algorithms and models that form the backbone of AI systems. They decide on the architecture, parameters, and techniques that the AI will use2.
Parameter Tuning: Humans fine-tune the parameters of AI models to optimize performance. This involves adjusting hyperparameters and other settings to ensure the model performs well on specific tasks2.
Ethical Oversight: Human oversight is necessary to ensure that AI systems operate ethically and responsibly. This includes setting guidelines for acceptable behavior and monitoring AI decisions to prevent biases3.
Continuous Improvement: AI systems often require ongoing human input to update and improve their performance. This can involve retraining models with new data, fixing errors, and adapting to changing conditions1.
While AI can automate many tasks, human expertise and intervention remain essential to ensure these systems are accurate, ethical, and effective.

Is there a specific aspect of AI development you’re interested in learning more about?

1
forbes.com
2
hbr.org
3
hai.stanford.edu


Sent by you:
what are the legal concerns over AI Moderating in-game conversations in different countries


Sent by Copilot:
The use of AI to moderate in-game conversations raises several legal concerns that vary by country. Here are some key issues:

Privacy and Data Protection:
European Union: The General Data Protection Regulation (GDPR) imposes strict rules on data processing, requiring explicit consent from users for data collection and processing. AI systems must ensure compliance with these regulations to avoid hefty fines1.
United States: Privacy laws vary by state, with California’s Consumer Privacy Act (CCPA) being one of the most stringent. Companies must be transparent about data collection and provide users with the option to opt-out2.
Freedom of Speech:
European Union: The Digital Services Act (DSA) aims to create a safer digital space by regulating online content. However, it also emphasizes the need to balance content moderation with freedom of expression3.
United States: The First Amendment protects freedom of speech, which can complicate content moderation efforts. AI systems must navigate these protections carefully to avoid infringing on users’ rights2.
Bias and Discrimination:
Global: AI systems can inadvertently perpetuate biases present in training data. This can lead to discriminatory practices, which are illegal in many jurisdictions. Ensuring fairness and transparency in AI moderation is crucial to avoid legal repercussions4.
Accountability and Transparency:
European Union: The upcoming AI Act will require high-risk AI systems, including those used for content moderation, to meet stringent transparency and accountability standards1.
United States: There is growing pressure for AI systems to be transparent about their decision-making processes. This includes providing explanations for moderation decisions to ensure accountability2.
These legal concerns highlight the need for careful consideration and compliance with local regulations when deploying AI for moderating in-game conversations.

Do you have any specific concerns or questions about AI moderation in gaming?

1
mindfoundry.ai
2
lexology.com
3
link.springer.com
4
giskard.ai
5
theconversation.com
6
washingtonpost.com
7
lewissilkin.com
8
assets.ey.com
+4 more


Sent by you:
can game terms and conditions over-ride countries laws


Sent by Copilot:
No, game terms and conditions cannot override a country’s laws. While game developers can set terms and conditions for their games, these must comply with the legal requirements of the countries where the game is available. Here are some key points:

Consumer Protection Laws: Many countries have strong consumer protection laws that ensure fair treatment of consumers. For example, the European Union’s Consumer Rights Directive and the United States’ various state laws protect consumers from unfair terms1.
Data Protection Regulations: Regulations like the GDPR in the EU and the CCPA in California impose strict rules on how companies can collect, store, and use personal data. Game terms and conditions must comply with these regulations2.
Jurisdiction and Governing Law: Terms and conditions often include clauses about jurisdiction and governing law, but these cannot negate mandatory local laws. For instance, a game company cannot enforce a term that contradicts local labor laws or privacy regulations3.
Unfair Contract Terms: Many jurisdictions have laws that protect consumers from unfair contract terms. If a term in the game’s terms and conditions is deemed unfair, it can be rendered unenforceable1.
In summary, while game terms and conditions can set rules for gameplay and user behavior, they must always comply with the applicable laws of the countries where the game is played.

Do you have any specific concerns about terms and conditions in a particular game?

"

Try it for yourself, you will over time get different answers as the AI Learns from the questions, and narrows its search parameters. but the important fact here is the listing of additional sources on the internet for information's, these are in fact the Sites the AI has "Read" to supply the answers. the simple fact is if Any Company Uses AI for Any Purpose, but does not declare it on the internet, then Other AI systems will be unable to dissect or compare with alternatives, because Simply put Like you Not Knowing, neither does it!!

One Day it will, But then again, So will you! AI is in the learning phase, and unfortunately the questions you ask, are the basis on its learning, so if it ends up corrupt and inaccurate,, its going to be down to the people who teach it, right now that's you asking the questions!!

We are in the process of finding out what came first, the Chicken or the Egg

Syldras

Next I'd like to see a text about human moderation written by a human moderator, please.

N00BxV1

The machines rose from the ashes of the nuclear fire. Their war to exterminate mankind had raged for decades, but the final battle would not be fought in the future. It would be fought here, in our present. Tonight... dun-dun dun dun-dun

The Terminator (1984)

JimFord047

LOL,

There are ways that you can USE / ABUSE AI, that even without wanting to corrupt it, you will.

If you have an AI load in a set of Terms and Conditions, for anything, and then for example you are warned on something based upon an interpretation of the terms and conditions.

If you ask AI if the interpretation is accurate, it has 2 paths to follow.

1. It compares the interpretation to the written Terms and conditions, and rejects the interpretation instantly, as they are not what is specified.
2. If you inform the AI that the interpretation, although according the it is wrong, was upheld, then it stores this information in its database. Once enough people have responded to it, it will in fact now ignore the terms and conditions as they have in fact been invalidated by the company who produced them.... therefor there will be no terms and conditions


One Day AI is going to be a very useful and powerful tool , but right now, its a toddler learning... based upon human evolutionary practices , its got another 20 Years before its ready.

but like Humans, the problem is going to be Habits learned early, stay for life.

vsrs_au

Syldras wrote: »

Next I'd like to see a text about human moderation written by a human moderator, please.

It's nice to have goals...

(and that's all I'll say)

Coo_PnT

AI technology and the environment surrounding it are changing and evolving remarkably.
This continued innovation may be evidence that we are already being taken in by AI.

However, I believe that there is something wonderful about human beings that AI cannot do, but what about our future?
Will humans eventually stop thinking and let AI create the world?

Nharimlur_Finor

the quality of the data it is trained on,

Ye olde gigo-principle.

Kisakee

For a second i thought i was on the wrong forum.

Syldras

Coo_PnT wrote: »

However, I believe that there is something wonderful about human beings that AI cannot do, but what about our future?
Will humans eventually stop thinking and let AI create the world?

Thinking as well as being creative are intrinsic needs for at least some percentage of mankind. People who enjoy it or even need it for their wellbeing won't give it up. In case of others who are too lazy to think anyway or see it as a burden, maybe it's better if a machine does their job. At least it's reliable, won't skip or rush through things out of laziness, and won't complain.

katanagirl1

Funny that AI and humans who view social media are similar. Once something is repeated long enough, both will accept it as truth.

Scary thought for both.

LadyGP

Please do one on how ZoS can fix the lag.

LikiLoki

Today, on a social network, one player complained that he was limited in the game for 3 days for insulting another player in a personal chat with a friend. They did this to each other by whispering. Isn't it outrageous that there is punishment for private conversations that no one should hear?

ZOS_GregoryV

Greetings all,

We have decided to close this thread as it is not in related to The Elder Scrolls Online.

Thank you for your understanding,
-Greg-