Are we creating goD...

"I can't think of many administrative functions of government that Ai could not drastically improve. From HR to logistics and planning, I'm thinking Ai wins every time. Right?"

On issues like this, I agree 100%.

But overall governance by AI .... no ways.

In my mind, i t would work best when integrated with human oversight to balance objectivity with empathy and ethical considerations

Just imagine if our submarines or other Navy vessels were all run or operating on AI

If that was the case, then the third world war , which would have been a nuclear catastrophy.... would have started in October 1962.

Rooinek is ay to understand thdifference between Government and State.   In  Democracies the people vote for their representatives to overn over them and is supposed to serve the voting public.    Aside from that there are two branches of Government that makes up the State - naely an independent judiciary who must ensure that the Constittitio of yj country is adhered to and the Publi Service bureaucrats that must execute the laws passed by the elected representatives of the people,  

his commence show that he does not realize how democracy functions - and he wants to be G overned by a system contra to democratic principles.   he AI system can improve administration - but they do what is provided to them very efficently - but the dnge is i can be fed information that will destroy democracy.   AI removing the lawmaking function or taking  it over destroy that part of democacy.    AI cannot take over the Judiciary either,   

What AI could do is to clean out the adminitrative and bureaucratic mess at oresent in use - but ny fuurther functioning  on legislative level wuld destroy democracy - but th en at one stage Rooinek rite on site he is no problem with teh Commnist system of governance - so one should not be surprised by the idiocy he wrote above,   

Rooinek is ay to understand thdifference between Government and State.   In  Democracies the people vote for their representatives to overn over them and is supposed to serve the voting public.    Aside from that there are two branches of Government that makes up the State - naely an independent judiciary who must ensure that the Constittitio of yj country is adhered to and the Publi Service bureaucrats that must execute the laws passed by the elected representatives of the people,  

his commence show that he does not realize how democracy functions - and he wants to be G overned by a system contra to democratic principles.   he AI system can improve administration - but they do what is provided to them very efficently - but the dnge is i can be fed information that will destroy democracy.   AI removing the lawmaking function or taking  it over destroy that part of democacy.    AI cannot take over the Judiciary either,   

What AI could do is to clean out the adminitrative and bureaucratic mess at oresent in use - but ny fuurther functioning  on legislative level wuld destroy democracy - but th en at one stage Rooinek rite on site he is no problem with teh Commnist system of governance - so one should not be surprised by the idiocy he wrote above,   

Okay, DA...I'll show you a trick that we used when we made our GPTs for the App.

Assume we had some sort of problem, wanted data presented in a certain way or wanted to avoid some undesirable outcome. We would provide ChatGPT with a clear description of the problem...

Let's assume that your post lays out the problems that a world governance AI would have to remedy before you could trust it to run the world.

Here is the command I typed into ChatGPT before posting your explanation for the problem...

User -> I wish to create a global governance AI. The goal is to be able to streamline governmental efficiency, share and reorganise data as necessary, and provide the citizenry with a  option to partake in referendum to promote direct governance by the population as much as possible. below is a summary of the general problem the population feel will be encountered by attempting to implement a global AI government. (I now pasted your post here).

Point out a) where the user is wrong and b) where the user has valid points and how we could adjust the AI to avoid these problems, or render them as non-issues.

Also take into account the potential for AI to become far more powerful in the coming years, and thus improving in it ability to prevent problems. You may also provide a list of instructions that the AI that would be run on to prevent particular problems arising.

ChatGPT --> 

1. AI Reflecting Societal Biases:

   • While it's true that AI is trained on existing data, advanced AI systems can be trained with algorithms that actively detect, counteract, and correct societal biases. Bias mitigation is a major focus of AI ethics research.
   • AI could potentially reduce biases present in human governance systems by providing decisions based on objective data, provided proper safeguards are implemented.

2. AI Lacks Empathy or Moral Reasoning:

   • Empathy and moral reasoning are not inherent to AI, but moral reasoning can be encoded in the system through ethical frameworks developed by diverse stakeholders.
   • AI can complement human decision-making by bringing rational analysis to situations where human emotion might cloud judgment.

3. Dependence on AI Reducing Human Oversight:

   • AI governance does not necessarily imply complete automation. Hybrid systems with human oversight can ensure critical thinking remains integral.
   • Structured fallback systems can engage humans during critical decisions or system malfunctions.

4. Failure of AI Leading to Government Collapse:

   • AI systems can be designed with robust redundancy and fail-safe mechanisms, including distributed governance structures that ensure continuity during system failures.

5. AI Working Better for the Technologically Savvy:

   • AI can be designed with highly accessible interfaces and mechanisms to ensure inclusivity, even for those who are not technologically adept.

6. AI Lacking Adaptability to Nuances:

   • Advanced AI systems are becoming increasingly adaptable and capable of nuanced decision-making through reinforcement learning and emergent properties. AI development is trending toward greater contextual understanding.

1. Accountability for Errors:

   • AI systems should operate under a framework where accountability is shared between human supervisors and system creators. Decision logs and transparent algorithms can clarify the chain of responsibility.

2. Surveillance, Privacy, and Data Misuse:

   • Safeguards such as decentralized data storage, encryption, and strict data governance policies are essential. Citizen consent and anonymization protocols should be foundational.

3. Ethical Considerations and Human Suffering:

   • Ethical principles should be embedded in the AI’s decision-making algorithms, with oversight committees to ensure alignment with societal values.

4. Vulnerability to Hacking and Manipulation:

   • Advanced cybersecurity measures, regular audits, and redundant systems should be integral to AI governance frameworks.

5. Cultural, Religious, and Social Sensitivities:

   • Diverse datasets and inclusive programming teams are critical. Regular consultations with cultural and community leaders should inform the AI's parameters.

6. Citizen Disconnect from Decision-Making:

   • Public interfaces for participatory governance, such as real-time referendums or feedback mechanisms, can maintain citizen engagement.

7. Creators Embedding Bias:

• AI governance systems must be developed transparently, with global oversight bodies to audit and evaluate fairness and neutrality in AI training.

Accountability for Errors:

"AI systems should operate under a framework where accountability is shared between human supervisors and system creators. Decision logs and transparent algorithms can clarify the chain of responsibility."

• Define Roles and Responsibilities:
  • Clearly delineate who is responsible for each aspect of the AI system: creators, developers, operators, and supervisors.
  • Assign accountability tiers (e.g., data accountability, algorithmic accountability, operational accountability).
• Legally Bind System Creators and Operators:
  • Require system developers and vendors to adhere to strict legal frameworks ensuring accountability for flaws, bias, or harm caused by the AI.

• Maintain Transparent Logs:

  • Record every decision made by the AI system, including:
    • Input data
    • Algorithmic process
    • Outputs and their justifications

• Enable Real-Time Monitoring:

  • Supervisors should have access to real-time logs, particularly for decisions involving critical or high-stakes situations.

• Auditable Records:

  • Ensure logs are tamper-proof and accessible for internal and external audits.

• Open Algorithm Policies:

  • Publish simplified explanations of AI decision-making processes in understandable terms for citizens and stakeholders.

• Third-Party Review:

  • Require independent experts to review algorithms for fairness, bias, and alignment with ethical standards.

• Explainability Standards:

  • Incorporate explainability into AI systems so that decisions can be comprehended by humans, especially in critical domains like law enforcement or healthcare.

• Human Oversight Panels:

  • Form independent panels comprising experts, stakeholders, and citizen representatives to oversee AI decisions.

• Error Review Protocols:

  • Develop standardized processes for reviewing, addressing, and rectifying errors made by AI systems.

• Classify Risks:
  • Categorize potential errors based on their severity (e.g., low, medium, high risk).
• Human Intervention Points:
  • Designate specific scenarios where human supervisors must intervene (e.g., life-and-death decisions, ethical gray areas).

• Liability Assignment:
  • Allocate liability proportionally:
    • Developers are accountable for coding errors or training flaws.
    • Operators are accountable for misuse or lack of supervision.
• Compensation Funds:
  • Establish a fund for victims of AI errors, funded by AI system developers or operators.

7. Education and Training for Supervisors

• Skill Development:
  • Train supervisors to interpret AI decisions and intervene effectively.
• Ethics and Compliance Training:
  • Educate on the ethical implications of AI governance and the importance of maintaining accountability.

8. Public Engagement and Transparency

• Citizen Reporting Mechanisms:

  • Implement channels for citizens to report errors or concerns about AI governance decisions.

• Transparency Reports:

  • Regularly publish reports detailing AI system performance, errors, and corrective measures.

9. Redundancy and Fail-Safe Mechanisms

• Multiple Oversight Layers:
  • Use layered systems where human supervisors, AI auditing systems, and oversight committees work together.
• Fallback Protocols:
  • Ensure contingency plans are in place to address AI errors quickly, minimizing harm.

10. Encourage Adaptive Learning and Improvement

• Continuous Learning Systems:

  • Design AI systems to learn from past errors and integrate solutions into their algorithms.

• Feedback Loops:

• Use citizen and stakeholder feedback to refine AI behavior and improve accountability mechanisms.
  
  

• Classify Risks:
  • Categorize potential errors based on their severity (e.g., low, medium, high risk).
• Human Intervention Points:
  • Designate specific scenarios where human supervisors must intervene (e.g., life-and-death decisions, ethical gray areas).

Detailed Strategy, Contingency Planning, and Testing for Risk Categorization and Mitigation

Objective:
To systematically identify, classify, and address potential risks in AI governance by their severity and impact, ensuring appropriate levels of response and human intervention.

1.1. Categorize Potential Errors

1. Low-Risk Errors:

   • Definition: Errors with minor consequences, such as inefficiencies or cosmetic issues.
   • Examples:
     • Slightly inaccurate statistical predictions.
     • Incorrect formatting in reports.
   • Response:
     • Automatically logged for system improvements.
     • Addressed during routine updates.

2. Medium-Risk Errors:

   • Definition: Errors that affect operational efficiency or minor stakeholder dissatisfaction.
   • Examples:
     • Misclassification of non-critical data.
     • Erroneous resource allocation.
   • Response:
     • Escalated to human review for corrective action.
     • Added to system training data for future mitigation.

3. High-Risk Errors:

   • Definition: Errors with significant societal, ethical, or life-threatening implications.
   • Examples:
     • Misjudgments in critical medical, safety, or legal decisions.
     • Ethical violations, such as biased rulings in a sensitive case.
   • Response:
     • Immediate escalation to human supervisors.
     • Comprehensive investigation and public reporting.

1.2. Designate Human Intervention Points

1. Scenarios Requiring Human Oversight:

   • Life-and-Death Decisions: Medical resource allocation, emergency disaster response.
   • Ethical Gray Areas: Culturally sensitive rulings, decisions involving conflicting rights.
   • Critical Public Infrastructure: Energy distribution, defense, or financial policy decisions.
   • Ambiguities in Data: When data quality is insufficient for AI to make an informed decision.

2. Triggers for Human Intervention:

   • Predefined thresholds for uncertainty in AI predictions (e.g., confidence below 70%).
   • Detection of ethical or legal conflicts flagged by the AI.
   • Citizen or stakeholder complaints about specific decisions.

2. Risk Mitigation Strategy

1. Preventative Measures:

   • Regular audits of training datasets for bias and inaccuracies.
   • Deployment of robust algorithms to flag and address low-confidence decisions.
   • Continuous updates to include new ethical standards, regulations, and cultural sensitivities.

2. Real-Time Monitoring:

   • Implement dashboards for supervisors to track AI decisions and flag anomalies.
   • Use predictive analytics to anticipate high-risk scenarios and allocate human resources.

3. Escalation Protocols:

   • Clearly defined pathways for escalating decisions to human supervisors.
   • Redundant systems to ensure decisions are not delayed in critical scenarios.

3. Contingency Planning

1. Error Detection and Response:

   • Immediate Lockdown: For high-risk errors, halt the AI system's actions in affected areas.
   • Fallback Mechanisms: Pre-configured human-driven workflows to replace AI decisions temporarily.
   • Incident Response Teams: Specialists ready to assess and mitigate high-risk errors.

2. Public Communication:

   • Transparent disclosure of errors and corrective measures.
   • Open channels for citizen feedback and reassurance to maintain trust.

3. Continuous Improvement:

   • Integrate lessons learned from errors into system updates.
   • Refine intervention thresholds and escalation protocols based on error analysis.

4. Methods for Testing

1. Simulated Risk Scenarios:

   • Develop real-world-like scenarios to test AI decision-making across all risk levels.
   • Examples:
     • Simulate an emergency response requiring life-and-death prioritization.
     • Test the AI’s handling of ambiguous ethical cases.

2. Stress Testing:

   • Push the system to its limits (e.g., high volumes of data, conflicting inputs) to evaluate performance and identify weak points.

3. Human Oversight Validation:

   • Conduct drills where supervisors intervene in simulated high-risk scenarios to ensure readiness.

4. Bias Detection Tests:

   • Analyze decisions for patterns of bias using diverse test datasets.

5. Redundancy and Fail-Safe Validation:

   • Test fallback mechanisms by deliberately shutting down components of the system.

Summary of Strategy

• Classify errors into low, medium, and high categories with clear examples and corresponding responses.

• Designate critical scenarios requiring human oversight, with clear thresholds and protocols for escalation.

• Employ proactive measures, real-time monitoring, and escalation procedures to reduce and address risks.

• Plan for error detection, response, and public communication, backed by a system for continuous learning.

• Use simulated scenarios, stress tests, and fail-safe validations to ensure the system is robust and adaptive.

This comprehensive approach ensures accountability, minimizes risks, and maintains public trust while leveraging the benefits of AI governance.

This took me about 10 minutes to produce. 

So, Ai is able to hold ethical, as well as technical, strategic, and other pertinent, information in its head all at the same time. 

So, using commercially available Ai today, we can already build a framework for governance that would outperform humans in efficiency, objectivity, planning. 

Of course there is separation between AIs built for a specific function, like monitoring ocean currents, and AI's built for governance of societies, they are not the same thing. The former being far more palatable to most people. 

But it will make sense for those niches AIs to stream into larger macro AIs that collate data and are able to use it to make strategic decisions that benefit humanity at scale. 

So, for example, the AI monitoring ocean currents feeds into a larger meteoritical AI, which itself feeds into emergency planning AIs, which in turn provide strategic advise to Emergency Services and government branches tasked with public safety, which in turn feed into other elements of governance. 

Here, it would seem illogical not to allow the AI to collate all this data, and then utilise its ability to strategize in order to minimise damage to property and people.

So while everybody agrees that AI could carry out administrative tasks, it's also true that governance is predominantly administrative. So where does AI get cut off and where do we insert the human in the decision chain above? Does the human just put the stamp on it?

And remember, we are only a the start of what AI is capable of. In the next few years its "ability" is going to go completely parabolic.

Excellent analogy Plum, also an interesting read..... thanks.

I do agree that AI is the way of the future, but I am not sure how well it will serve us in the long run.

As it is now, the internet, as well as AI, has made the human race lazy..... more effecient, and more effective.... but lazy.

Too lazy to think for ourselves, work out our own problems, and to sit and actually just "think"

Scholars have already become lazy, getting ChatGPT to do their assignments for them, getting answers to questions that could have taken someone 3 hours to find before, now gets provided in 2 seconds... and this just hinders the cri tical thinking, problem solving or even just independent thinking... as well as creative thoughts...never mind the authenticity or plagiarism on projects or thesis.

AI still needs to be programmed so it can generate convincing yet inaccurate information as well, and who knows what any algorithm would really involve.

I also believe in the long run that the overuse of AI will diminish our proficiency in writing, reasearch and analysis from a human stand point.

Don't get me wrong, medical and scientific advancements with AI is fantastic for the human race, but at what ultimate cost will it come, especially in other areas if it has total control.

The job displacement that comes with a fully operational AI society is also a huge concern, especially in industries like customer service, manufactoring and data entry 

I also think the heavy use of AI will reduce our social skills and affect inter-personal relationships...but again I am looking at the long term effects.

AI is fantastic, and it's here to stay, but to give it total control and governance over the people, no ways...... however I do see that even most of your examples above also refer to human oversight in most of the stages of AI use, which is what I am all for... just not complete control, because I still believe that would be very reckless. 

Pres

I think Democatic Governments are already looking forward to more efficient Governance - but not of AI taking over Governance,   Democrsatic Govenance relates to people selecting their representatives in the legislature - who are supposed to be accountable to the people they represent.   Any uneleced people taking  charge of AI represents the destruction of basic human rights and freedoms by a dicatorial group of people ho never eneds toa ccount for qhat they dio to people.              

Politicians almost without exception are corrupt self-serving liars.

It amazes me that anyone would choose these dregs of humanity rather than any alternative.

It is choice beteen democracy and dictatorial Government that remains the issue.  I knw you support the ultra-corrupt Biden Admsinistration - but only idiots will want to be governed by people that has the power to do what they please and no accountability to anyone but themselves.

The choice is  between dictartorships and democracy and my choice will always be democracy,  In the case of the Biden Presidency there was no accountabiliy because the FBI protected the criminals in Government - but that would stop soon since the majrity of the people decided on who should govern them.  

The AI system would make precise bugetting and budgetting control massively - but the fact uis thta budgt nee legislative approval.    The udgets cn be done in more detail and control mechanisms by AI will make corrupt practices impossible epcially if you deaaaaaaal with thist like the Bidens and other Democrat olitcians ho stole billion from the Government.   Take anotehr examin the USA.   No audits were done since 2003 of the USA Department of Defense abd in 2023 the Republicand and some of teir democratic collegues efuse to pass the DOD budget unless auditing statr immediately,   So the audting start and 11 projects were identified according to which the audit of the $520 billion budget was to be subject.   After the first 7 Projects was completed there was $80 billion spent on thinks nbody could provide infor on hat the money a used for,    Tat amount is likely to rise to $250 billion when all audits have been completed,  So what was that money used for?   Top of te list gross incompetence and gross bribery and corruption.  

AI can solve that problem and at the same time provid data that bureaucrats can be eld accountable - so that part is 100% acceptable - but for AI to decide onm what should eb in the budget proposals are BS,    

By the way there are  way more corrupt bureaucrats then tere are corrupt politcians.  

.       

This is quite a read, but I think it really is worth it if you have time....due to the interesting stats below from ChatGPT

Sorry if I don't get all the text aligned, I have tried to do so but the formatting was a bit all over the place.

I believe that AI use would be fantastic for a ny government ... because just the ease and efficiency of working with it every day would be phenomenal, and it would definitely speed up any daily government bureaucracy

No lunch breaks

No tea breaks

No bad attitude days

No leave days

No sick days

No theft

No dishonsty 

No biased service based on race

Trust me, I am all for going AI in certain aspects of running any organisation, province or country, just not total autonomy .... hell no.

However, when one looks at how this would negatively affect the jobs in just South Africa... never mind the US, China etc etc... where does one draw the line....

For instance, AI would be able to do the following:

1. Government and Administration

AI could handle tasks like:

• Public Service Management: Automating application processing for permits, IDs, and licenses.

Data Analysis: Identifying trends in public health, crime, or education to inform policies.

• Taxation: Streamlining tax collection and audits with AI-driven systems.

Policy Implementation Monitoring: Tracking and reporting on the progress of government programs.

2. Healthcare

AI can significantly contribute to:

• Diagnostics: Analyzing medical scans, test results, and patient data to identify conditions.

Administration: Managing patient records, scheduling, and resource allocation.

• Telemedicine: Offering virtual consultations and basic health advice.

Public Health Planning: Predicting and managing outbreaks using data models.

3. Law Enforcement and Justice

AI could take over:

·       Crime Prediction and Prevention: Analyzing patterns to allocate resources effectively.

 ·       Judiciary Tasks: Supporting judges by reviewing case precedents or handling minor disputes via automated mediation.

·       Surveillance: Monitoring CCTV for suspicious activities.

4. Education

AI systems might:

·       Personalize Learning: Adapting curricula to students' individual needs.

·       Administration: Automating grading, scheduling, and school resource management

·       Virtual Tutoring: Providing supplemental education through AI tutors.

5. Transportation and Infrastructure

AI could manage:

• Traffic Systems: Real-time traffic flow optimization and accident prevention.

Public Transport: Scheduling and route optimization for buses, trains, and other public systems.

• Infrastructure Maintenance: Predictive maintenance using IoT and AI analytics.

6. Social Services

AI could assist in:

• Benefit Distribution: Automating eligibility checks and fund transfers.

Case Management: Identifying families or individuals needing urgent intervention based on data.

• Disaster Response: Managing resource allocation during crises.

7. Agriculture

AI applications in farming include:

• Crop Monitoring: Using drones and AI to assess crop health and optimize yields.

Supply Chain Management: Predicting demand and managing logistics for agricultural products.

8. Financial Management

AI could take over:

• Budgeting: Optimizing national and provincial budgets using data insights.

Fraud Detection: Monitoring transactions to detect and prevent corruption.

----------------

I then asked AI how it would affect the job displacement in South Africa if it only took over the jobs that it was capable of performing.

----------------

1. Total Workforce and Government Jobs

• Government Employees: South Africa has approximately 1.3 million public sector employees.

Total Workforce: The labor force is about 15.2 million people as of 2023.

2. Sectors Where AI Could Take Over

AI's impact would vary across sectors:

Government and Public Administration

• Jobs Impacted: About 40-60% of government jobs (520,000–780,000 jobs).
•  
• Clerical roles, data management, and repetitive tasks could be automated.

High-level decision-making roles and positions requiring empathy or leadership would remain.

Healthcare

Jobs Impacted: Approximately 20-30% of healthcare jobs.

• Administrative and diagnostic roles are highly automatable.

Hands-on medical care (doctors, nurses) would see less impact.

Law Enforcement and Justice

• Jobs Impacted: Around 10-15% of the police force (19,000–29,000 jobs).
•  
• Tasks like  monitoring and report writing could be automated.

Education

• Jobs Impacted: About 30-40% of education-related jobs (120,000–160,000 jobs).
•  
• AI tutors and automated grading could replace some teaching assistants and administrative roles.

Transportation and Infrastructure

Jobs Impacted: Around 20-30% of related jobs.

• AI could optimize public transport and logistics, potentially affecting roles like dispatchers and maintenance planners.

Agriculture

Jobs Impacted: About 15-25% of agricultural jobs.

• Roles in crop monitoring and supply chain management are at risk, though manual labor will still be needed.

Social Services

• Jobs Impacted: Approximately 20-25% of roles.
•  
• Benefit distribution and case management could see significant automation.

Financial Services

Jobs Impacted: Around 30-40% of jobs in finance.

• Accountants, auditors, and administrative financial workers could be replaced by AI systems.

Estimated Total Jobs Affected

Based on these sectoral impacts, 30-40% of South Africa's workforce could be affected by AI, translating to 4.5 to 6 million jobs. These jobs would either be displaced, transformed, or augmented by AI.

Now just imagine the social and economic implications of these numbers, if almost 6 million South African people all lost their jobs ….say….within the next 2 years to AI... with more losing their jobs every single year as AI advances...

Now extrapolate that to all other countries around the world…

The implications are huge

"AI systems should operate under a framework where accountability is shared between human supervisors and system creators. Decision logs and transparent algorithms can clarify the chain of responsibility."

1...Who will these "supervisors" and "system creators" be? Who will appoint them, who will make sure they are not corrupt, self serving?

"Mike, you didn't read what I wrote.

Representative democracy is too easily corrupted.

We have the technology now that could mean that people represent themselves and vote on issues after being presented with unbiased and accurate information. They don't need some forked tongued twat to speak for them."

2...again, with tech becoming so sophisticated,  who will guarantee that it's not misused for nefarious purposes?...the higher the level of sophistication,  the more difficult it would be to be truly transparent...which makes oversight almost impossible...look at how it was already misused in social media and mainstream media.

"Politicians almost without exception are corrupt self-serving liars.

It amazes me that anyone would choose these dregs of humanity rather than any alternative."

3..The ones controlling these dregs ATM, will allso be controlling the AI...AI will be less prone to detection than the dregs though...you are advocating handing over of the keys of the kingdom to a selected few...elite individuals who have already proven that they can't be trusted...

PS...DA, I know we are at loggerheads sometimes,  but I'm 100% with you on this...some very good points.

BTW, I had a chat with my optometrist a few years back...they are already worried that 99% of their function will be taken over by AI and 3D printing technology...the technology are already there...

"again, with tech becoming so sophisticated,  who will guarantee that it's not misused for nefarious purposes?...the higher the level of sophistication,  the more difficult it would be to be truly transparent...which makes oversight almost impossible...look at how it was already misused in social media and mainstream media"

Spot on, Twitter and Facebook is a huge example

"The ones controlling these dregs ATM, will allso be controlling the AI...AI will be less prone to detection than the dregs though...you are advocating handing over of the keys of the kingdom to a selected few...elite individuals who have already proven that they can't be trusted..."

Exactl y

"PS...DA, I know we are at loggerheads sometimes,  but I'm 100% with you on this...some very good points."

I agree with your points as well Db, and even some of Plum's as well.... it's a very intriguing and thought provoking concept.

AI introduction to everyday life as we know it is clearly a very sophisticated and intricate scenario that has so many incredible positives for all of mankind..... whilst having so many potentially disastrous negatives, which is why I would always say that full autonomy is reckless and stupid.

Human oversight I can accept and I would fully endorse, but only for certain operational requirements and industries...

One also has to look at the world and ask if it is better off financially than it was 20 years ago...40 years ago.

See this US national debt chart...

That's the most powerful country on earth...

I mean, I don't know how you fix that. 

Notice how the truly parabolic rise started at around the same time that the war on terror did. And Covid sent it almost vertical.

China is in an even worse position btw.

Honestly, it looks to me like we should take any help we can get to normalise economies again. Politicians, not only in the US, but in many countries, are failing. The system is simply not working for a lot of people in the world. 

I could be wrong, but the US spent $6T on the Iraq and Afghanistan wars. That is 20%  of their current national debt, before interest and related costs. FFS!? An AI would not have gotten suckered into that. It would have looked at Powell's evidence and found it to be short of the mark by some distance. An AI would certainly not have had its vice president also benefiting from contracts relating to the war. 

We can speculate as to, and consider, hardships that AI would usher in as we assimilate, but it's also worth looking back and asking how much hardship could have been avoided had we extremely powerful AI at the time.

Assume that AI wasn't in charge, but that there was AI present when Colin Powell presented his evidence to the UN. And he has to try and bullshit the AI...he wouldn't get very far. 

The AI could evaluate evidence and motive, provide pertinent context, project costs and deliver pretty unbiased predictions as to outcomes. And it could do it right there, on the spot...no need to schedule a meeting for next week. And that AI could be open source and completely decentralised. A true representation of human interest. 

Had that been in place, there would have been no Iraq war, likely nothing in Afghanistan either, and the US would have less debt and a much better image on the world stage.

I'm not saying that AI should dictate how I live my life, but I am saying that it would be foolish to ignore its potential to help us make the correct decisions. One could probably say that the better the advice is, the more closely it resembles a type of control.