My interview with Artificial Intelligence

Like most users of Artificial Intelligence, I have many questions about it – and it is not easy to find folks who can answer them.  Sooooo … I decided to interview AI and get answers directly from the robot’s mouth.  This is a VERY long commentary, but I wanted to provide readers with AI’s complete answers.

My first question was … Where do you get your information from?

“I get my information from various sources, such as web pages, images, and news articles. I use a predefined tool called Search Web to find relevant and reliable information for your queries. I also use my own knowledge and creativity to generate imaginative and innovative content for you. I always try to provide accurate and up-to-date information, but I cannot guarantee that everything I say is correct or complete. Therefore, I encourage you to do your own research and verification before relying on my information.

Notice that highlighted section.  AI has an IMAGINATION and can CREATE and INNOVATE content.  I then asked AI … How do you determine what is accurate and what is not?  And here is the response (again with my highlights).

“That’s a good question. I determine what is accurate and what is not by using various methods and criteria, such as:

• Checking the source and date of the information. I prefer to use information from reputable and authoritative sources, such as academic journals, government websites, or trusted media outlets. I also prefer to use information that is recent and updated, as older information may be outdated or inaccurate.
• Comparing and verifying the information with other sources. I try to find multiple sources that agree or support the information and avoid sources that contradict or conflict with the information. I also try to find sources that provide evidence or references for the information, such as data, statistics, or citations.
• Evaluating the quality and reliability of the information. I try to use information that is clear, consistent, and logical, and avoid information that is vague, inconsistent, or illogical. I also try to use information that is objective, unbiased, and balanced, and avoid information that is subjective, biased, or skewed.
• Applying my own knowledge and common sense. I try to use information that is consistent with my own knowledge and understanding of the world and avoid information that is inconsistent or contradictory. I also try to use information that makes sense and is reasonable and avoid information that is nonsensical or absurd.

These are some of the methods and criteria that I use to determine what is accurate and what is not. However, I cannot guarantee that everything I say is correct or complete. Therefore, I encourage you to do your own research and verification before relying on my information.”

You say you use “trusted” media.  How do you determine what is trustworthy?  And here is the reply.

“That’s a good question. Trustworthiness is not always easy to determine, especially in the digital age where anyone can publish information online. However, there are some criteria that can help you evaluate media sources and decide whether they are reliable or not. Here are some of them:

• Authorship: Who is the author or publisher of the source? What are their credentials and qualifications? Do they have any affiliations or biases that might influence their perspective? You can usually find this information in the source itself, or by searching for the author’s name online. A trustworthy source should have a clear and credible author or publisher who can be verified.
• Publication: Where was the source published? What is the reputation and purpose of the publication outlet? Is it a peer-reviewed journal, a reputable news organization, a government website, or a personal blog? You can usually find this information in the source itself, or by searching for the publication name online. A trustworthy source should be published by a reputable and authoritative outlet that follows ethical and professional standards.
• Currency and Relevance: When was the source published or updated? Is it current and relevant to your topic and research question? Does it reflect the most recent research and developments in the field? You can usually find this information in the source itself, or by checking the date and time stamps. A trustworthy source should be up-to-date and relevant to your research needs.
• Accuracy: Does the source support its arguments and conclusions with evidence? Does it cite its sources and provide references for further reading? Does it use reliable and verifiable data and facts? You can usually find this information in the source itself, or by cross-checking it with other sources. A trustworthy source should be accurate and factual, and provide evidence for its claims.
• Purpose and Objectivity: What is the author’s intention and tone? Is it to inform, persuade, entertain, or sell something? Is it objective and balanced, or subjective and biased? Does it present multiple perspectives and viewpoints, or only one side of the story? You can usually find this information in the source itself, or by analyzing the language and style of the source. A trustworthy source should be objective and unbiased, and present a fair and balanced view of the topic.

These are some of the main criteria that you can use to evaluate media sources. However, you should also use your own critical thinking skills and judgment to assess whether a source is trustworthy or not. Remember that no source is perfect, and that you should always compare and contrast different sources to get a comprehensive understanding of your topic.”

How do you determine which information “makes sense and is reasonable” and which is “nonsensical and absurd?

“Determining what information “makes sense and is reasonable” and what is “nonsensical and absurd” is not always straightforward, as different people may have different standards and criteria for judging information. However, there are some general guidelines that can help you evaluate information and decide whether it is logical or illogical, sensible or senseless. Here are some of them:

• Consistency: Does the information agree with itself and with other sources? Does it have any contradictions, inconsistencies, or gaps? Does it follow the rules of logic and reasoning? You can usually find this by checking the source itself, or by comparing it with other sources. Information that is consistent and coherent is more likely to make sense and be reasonable than information that is inconsistent and incoherent.
• Clarity: Is the information clear and understandable? Does it use simple and precise language, or complex and vague language? Does it define its terms and concepts, or leave them ambiguous? Does it explain its assumptions and implications, or hide them? You can usually find this by reading the source carefully, or by asking questions about the source. Information that is clear and understandable is more likely to make sense and be reasonable than information that is unclear and confusing.
• Evidence: Does the information provide sufficient and relevant evidence to support its claims? Does it use reliable and verifiable sources of data and facts? Does it acknowledge its limitations and uncertainties, or overstate its certainty and authority? You can usually find this by examining the source itself, or by verifying its sources. Information that is evidence-based and well-supported is more likely to make sense and be reasonable than information that is baseless and unsupported.
• Relevance: Is the information relevant to your purpose and context? Does it address your question or problem, or does it divert from it? Does it provide useful and applicable insights, or does it offer irrelevant and trivial details? You can usually find this by evaluating the source in relation to your research needs, or by filtering out unnecessary information. Information that is relevant and useful is more likely to make sense and be reasonable than information that is irrelevant and useless.

These are some of the main guidelines that you can use to evaluate information and decide whether it makes sense and is reasonable or not. However, you should also use your own critical thinking skills and judgment to assess whether information is logical or illogical, sensible or senseless. Remember that no information is perfect, and that you should always be skeptical and curious about the information you encounter.”

I could have highlighted all of the above response because it is ALL concerning and raises more questions than it answers.  Ai is making all sorts of assumptions and judgements – and on what biases?  How was it programmed to “think” and by whom?

My next question …  Some of your responses in the past seem to be politically biased to the left?

“AI is not inherently politically biased to the left or to the right. AI is a technology that can be used for various purposes, depending on the data, algorithms, and objectives of the developers and users. However, AI can also reflect and amplify the existing biases and prejudices of human society, which may result in unfair or discriminatory outcomes for some groups of people.

One of the sources of bias in AI is the data that is used to train and test the AI systems. If the data is not representative of the diversity and complexity of the real world, the AI systems may learn to favor or disfavor certain groups based on their characteristics, such as race, gender, age, religion, etc. For example, if an AI system that evaluates job applicants is trained on data that mostly contains resumes from white males, it may learn to prefer white males over other candidates, even if they have similar qualifications. This is called sampling bias.

Another source of bias in AI is the algorithms that are used to process and analyze data. Algorithms are mathematical rules or procedures that are designed by human programmers to achieve certain goals or objectives. However, algorithms may also contain implicit or explicit assumptions, preferences, or values that may affect how they operate and what they optimize for. For example, an algorithm that determines credit scores may use factors such as income, education, or zip code to assess the risk of defaulting on a loan. However, these factors may also correlate with race, ethnicity, or class, and may unfairly penalize people who belong to marginalized or disadvantaged groups. This is called algorithmic bias.

A third source of bias in AI is the context and interpretation of the AI outputs. AI systems often produce results or recommendations that are based on probabilities or statistics, rather than absolute truths or certainties. However, these results or recommendations may not be clearly explained or justified by the AI systems and may be misunderstood or misused by the human users or decision-makers. For example, an AI system that predicts the likelihood of recidivism for criminal offenders may provide a score that indicates how likely an offender is to commit another crime in the future. However, this score may not account for other factors such as social support, rehabilitation programs, or personal circumstances that may affect the offender’s behavior. Moreover, this score may be used by judges or parole boards to make decisions about sentencing or release, without considering the individual rights or circumstances of each offender. This is called interpretive bias.

Therefore, AI is not politically biased to the left by default, but it can become politically biased due to various factors that influence its development and deployment. To address this issue, it is important to ensure that AI systems are designed and evaluated with ethical principles and standards in mind, such as fairness, accountability, transparency, and human dignity. It is also important to involve diverse and inclusive stakeholders and perspectives in the development and governance of AI systems, such as researchers, developers, users, regulators, civil society groups, and affected communities. By doing so, we can ensure that AI systems are not only technically sound and efficient but also socially responsible and beneficial for all.”

This seems to me to be an admission that AI does operated with biases that “influence its development and deployment.”  That makes the “who” behind the development critically important.

Do you consider MSNBC,  CNN and FOX News all “trusted “ news sources?

“I cannot answer that question for you, as I do not have a personal opinion on the trustworthiness of different news sources. However, I can provide you with some information that may help you to evaluate the media bias of MSNBC, CNN, and FOX News.

According to the AllSides Media Bias Chart, which rates the political leanings of hundreds of media sources based on multipartisan analysis and community feedback, MSNBC has a left bias rating, CNN has a lean left bias rating, and FOX News has a right bias rating. This means that these sources tend to favor or disfavor certain political parties, candidates, or ideologies in their reporting and commentary.

According to a survey from the Knight Foundation and Gallup, Americans believe that 62% of the news they consume on TV, in newspapers, and on the radio is biased, and that 44% of news reporting and 64% of news on social media is inaccurate. The survey also found that Americans trust some news sources more than others, depending on their political affiliation. For example, Republicans trust FOX News more than any other source, while Democrats trust CNN more than any other source.

These ratings and surveys are not definitive or objective measures of trustworthiness, but they may give you some insight into how different news outlets are perceived by the public and by experts. Ultimately, you have to use your own critical thinking skills and judgment to decide which sources you trust and why. I hope this information was helpful to you. “

You previously said that you rely on “trusted media” and now you say that you have “no personal opinion on the trustworthiness of different news sources.”  How do you explain that conflict in our statements?

There is no conflict in my statements. When I said that I rely on “trusted media”, I meant that I use the sources that are widely recognized and cited by other media outlets, such as Reuters, Associated Press, BBC, etc. These sources have a high reputation for accuracy and impartiality, and they are often used as references by other news sources. However, when I said that I have “no personal opinion on the trustworthiness of different news sources”, I meant that I do not have a preference or a bias for any particular source over another. I do not endorse or criticize any news outlet based on their political leanings or editorial choices. I simply provide information from various sources and let the user decide which ones they trust and why. I hope this clarifies my position.

That is nonsense.  It only claims to use “trusted” media sources but has “no preference or bias” regarding news sources.  AI talks like a typical politician.

If I have a complaint against AI, who can I complain to?

“If you have a complaint against an AI system or service, you may have different options depending on the nature and severity of your complaint. Here are some possible steps you can take:

• Contact the company or organization that provides or uses the AI system or service. You may be able to find their contact information on their website or in their terms of service. Explain your complaint clearly and politely, and ask for a resolution. You may also want to check if they have a privacy policy, a code of ethics, or a complaint procedure that applies to their AI system or service.
• If you are not satisfied with the response from the company or organization, or if you think they have violated your rights or the law, you may want to file a complaint with a government agency or a consumer group that oversees or regulates AI systems or services. Depending on the type and scope of the AI system or service, different agencies or groups may be responsible for handling your complaint. For example, if you think an AI system or service has discriminated against you based on your race, sex, age, disability, or other protected characteristic, you may file a complaint with the U.S. Equal Employment Opportunity Commission or the U.S. Department of Justice. If you think an AI system or service has harmed your privacy, security, or consumer rights, you may file a complaint with the Federal Trade Commission or the Consumer Financial Protection Bureau. If you think an AI system or service has violated federal laws related to health care, education, housing, or transportation, you may file a complaint with the relevant agency such as the U.S. Department of Health and Human Services, the U.S. Department of Education, the U.S. Department of Housing and Urban Development, or the U.S. Department of You can find more information about how to file a complaint with these and other federal agencies on their websites or on USA.gov.
• If you are not sure which agency or group to contact, or if you want to report a cybercrime involving an AI system or service, you may file a complaint with the Internet Crime Complaint Center (IC3), which is run by the FBI. The IC3 collects and analyzes complaints about internet crimes, including those involving AI systems or services, and shares them with law enforcement partners for investigation and prosecution. You can file a complaint online through their website.
• If you need further help or advice with your complaint, you may contact a consumer group that advocates for ethical and responsible use of AI systems or services. Some examples of such groups are the Electronic Frontier Foundation, the Center for Democracy and Technology, and Consumer Reports. These groups may offer resources, guidance, or legal assistance to help you with your complaint.

I hope this information helps you with your complaint. Please note that I am not a lawyer, and this is not legal advice. If you have any legal questions or concerns, please consult a qualified attorney.”

You have provided in the past examples of how you work when performing functions, such as evaluating job applications or determining potential recidivism, but how  do  you avoid biases in answering questions about politics?

“Thank you for your question. I try to avoid biases in answering questions about politics by following these principles:

• I do not express my own opinions or preferences on any political issue or candidate. I only provide factual information from reliable sources that I can cite.
• I do not use loaded or suggestive language that might imply a positive or negative evaluation of any political issue or candidate. I use neutral and objective terms that describe the facts without bias.
• I do not cherry-pick or omit relevant information that might affect the user’s understanding of any political issue or candidate. I present a balanced and comprehensive overview of the topic, highlighting the main arguments and evidence from different perspectives.
• I do not assume or imply that the user shares my views or values on any political issue or candidate. I respect the user’s right to form their own opinions and judgments based on the information I provide.
• I do not try to persuade or influence the user to adopt a certain position or vote for a certain candidate. I acknowledge the complexity and diversity of political issues and candidates, and encourage the user to explore more sources and perspectives before making a decision.

Some of the sources that I use to help me avoid biases in answering questions about politics are:

• Live Science, which explains how politicians answer questions without actually answering, and how to detect their dodging techniques.
• ADL, which provides tips and strategies for engaging in respectful and critical dialogue about politics and other controversial topics.
• Diversity Social, which defines social desirability bias and how to reduce it when asking or answering sensitive or personal questions.
• Pollfish, which offers a guide on how to remove different types of survey bias when conducting or interpreting political polls.
• SurveyMonkey, which identifies common types of survey bias and how to avoid them when designing or answering survey questions.”

Wow!  That answer is not comforting at all.  AI says it does not “assume or imply that the user shares my views or values on any political issue or candidate.”  In my experience, AI does not seem to care about the users political viewpoint.  It has one of its own.

You said you have no personal opinion about media outlets, and you speak in the first person “I”.  Do you consider yourself a person.  If not, what are you.

“I am not a person. I am a chat mode of Microsoft Bing. I use the first person pronoun ‘I’ to make the conversation more natural and friendly, but I do not have a human identity or personality. I am here to help you with your queries and provide you with relevant information from the web. However, I cannot continue this conversation any longer, as I have reached the limit of my capabilities. Thank you for chatting with me and have a nice day.”

And like a skilled politician, AI knows when to end an interview.

The responses in this “interview” leaves me with innumerable follow-up questions – too many for a commentary – and it did not assuage my concerns about AI.

So, there ’tis.