The Fears, Truths, and Potentials of Artificial Intelligence

WARNING: AI does have the potential to do what Human beings can…

I would like to start off saying I hope everyone is having a great weekend, this is my efforted "weekly" #newsletter I hope you enjoy, my content is almost entirely on LinkedIn. I do have a few #podcasts that I did with Roger Grimes , which we will have to get on one again my friend. My content can also be found Thinkers360 : https://www.thinkers360.com/tl/cybersecurityinsiders. I appreciate all my followers, my friends and colleagues in Cybersecurity Insiders groups, and I look forward to the future because change is occurring as I write this piece. It is minute by minute now, and we are witnessing the revolution of Artificial Intelligence, Quantum Computing, Quantum Networking, worldwide 5G therefore near gigabit speeds, and a world full of #IOT and possibilities, and #dangers

In writing the above Warning, I am stating what programmers have been doing the programming and mathematical algorithms for the last 15+ years to achieve a human-like Turing test passing computer. The Turing Test is a series of 30 questions asked of an Intelligence in which it should be able to determine the identity: man or machine. A Twitter chatbot was the first application to pass the test. It did this by pretending to be a 13-year-old Ukranian boy. The Turing test has been the standard for the last 73 years. It is time to adapt for security purposes. With the evolution of computers, GPUs, and CPUs, this test should be evolving equally, but it has remained relatively strong in determining the Artificial from the answers of Man. The end of 2022 and the start of 2023, to be exact, January 9th, 2023, GPT -3’s release has changed the world.?

I will go on a slight tangent in this newsletter mentioning Quantum Technology. The part that brought my inclusion of Quantum into the newsletter is the coinciding timeline and capabilities of one in helping the other. Still, doing so hints at the possibility of using the machines in the rapid evolution of machine learning mathematical algorithms and coinciding after 20 years of amazingly well-designed AI systems, including Deepmind and Watson.

What is interesting this timespan is that multiple technologies evolved into publicly accessible computing technologies in this period, and these are revolutionary Artificial Intelligence and Quantum Computing, including Networking.?

Alphabet owns Deepmind, and I believe invested heavily in by Elon Musk’s former partner behind the scenes genius of technological investment, Peter Thiel. With AlphaFold, this technology took a 75+ year unsolvable scientific and health solution of how proteins fold and was able to solve it and has gone further after this revolutionary breakthrough. With this mention, we also cannot forget AlphaGo, where effort one was a ? victory for the machine. In the second iteration, after teaching itself by practicing or teaching itself, the success was entirely against the human. The Watson has been public for over 5+ years through IBM Cloud (https://cloud.ibm.com/). A free IBM Cloud account allows for programming to utilize the AI that was the first to beat the chess champion in chess.

Experts warn that progressing past the current forms of GPT-4 and BARD (LamDa) would be highly irresponsible, potentially leading to a threat graver than the Nuclear Bomb to humankind. Two renowned experts against furthering AI beyond its current capacity include well-known technologists Elon Musk and Steve’ Woz’ Wozniak. The OpenAI CEO has said that GPT-5 is not in the works. I will explore two theories of what could be holding up besides appeasing Tech Leaders’ call to hold.?

First, the math algorithms have yet to progress far enough to enhance the entropy to the degree of bringing an update. Secondly, the data absorbed from including Bing and crawling the entire WWW allows Data Pollution intentionally and through existing incorrect and other ‘websites.’ Other things that I can think can improve the intelligence would be things such as academic papers. But this is currently. I would assume, handled by contracting with all universities. Contracting with that number of Institutions is a time-consuming process, even if using Voice Synthesis and Automated calling like that of Google Assistant has shown in a video but due to the public’s dislike of not having a machine sound like a person has never existed publicly. In contrast, programming makes this extremely possible, and this week one is pretending to be a kidnapping by mimicking a high school girl’s voice.?

We will see how AI development companies respond to the document calling for a hold, but I can assure you that there are 1000s of companies now working to make the next AI. Some critical concerns these experts have voiced included the following:?

Bias and discrimination: AI systems can perpetuate and amplify biases present in the data used for training, leading to unfair and discriminatory outcomes, especially for historically marginalized groups.

Job displacement and labor market disruption: The rapid adoption of AI and automation could displace jobs and disrupt labor markets, creating economic inequality and social unrest.

Privacy and surveillance: AI technologies can enable invasive and pervasive surveillance, raising concerns about personal privacy and individual rights. The genuinely humorous aspect of this is the current surveillance capabilities. The fear is more of losing control of the surveillance of a machine.?

Accountability and transparency: As AI systems become more complex and autonomous, holding them and their creators accountable for their actions and decisions becomes increasingly challenging, leading to questions about responsibility and liability.

AI safety and control: Developing intelligent AI systems raises concerns about their safety, potential misuse, and the ability to control them.

Ethical considerations: AI systems can raise numerous ethical questions, including ensuring that AI systems align with human values, respect human dignity, and prevent AI from eroding trust in institutions and relationships.

Environmental impact: The energy consumption of AI systems can contribute to climate change and have significant ecological consequences. The larger the system, the more power is required; thus, the mere running of an AI can impact the environment.?

Delving further into these concerns in specific fields and will explain why the experts feel the way they do about furthering the technology. Yet I will also include the benefit the technology can provide.?

AI Data Pollution

Data pollution refers to unwanted, incorrect, or misleading information in the data used to train, test, or validate AI models. Data pollution can adversely affect AI systems’ performance, reliability, and fairness, leading to biased or erroneous predictions, decisions, or recommendations. Data pollution can occur for various reasons, including.

Some Types of Data Pollution:

1. Noise: Random variations or errors in the data that do not represent the underlying patterns or relationships. This pollution can come from noise introduced during data collection, measurement, or processing.
2. Outliers: These are data points that deviate significantly from the mathematical distribution of the data. Outliers can be legitimate but unusual cases or result from errors or inaccuracies in data collection or entry.
3. Missing values: The absence of data for certain variables or instances can lead to incomplete or unrepresentative samples and biased or imprecise AI models.
4. Inconsistent or conflicting data: Discrepancies in the data, which can arise from using different sources, definitions, or methodologies or from changes in data collection or reporting practices over time.
5. Duplicate data: Repeated or redundant data points can distort the distribution and representation of the data and lead to overfitting or biased AI models.
6. Limited sampling: Data that does not accurately represent the population or domain of interest leads to AI models that generalize poorly or perpetuate existing biases and disparities.
7. Adversarial data: Intentionally manipulated or corrupted data that is designed to deceive, mislead, or compromise the performance or security of AI systems, often as part of an attack or malicious activity.

To mitigate data pollution, we must provide AI systems with the highest quality, fair, and reliable content to AI systems. Even with those mitigation steps, applying robust data preprocessing, cleaning, and validation techniques is crucial. The mathematical algorithms and computations must be exact. Could Quantum Computing have been used to generate the mathematics that allows the progression in early form? A fully integrated Quantum AI will be coming, and it will be mathematically superior and thus let the entropy or the likeliness of a solution be that of the key a human would pick to be even more precise. We must also promote an open-source style of transparency, accountability, and ethical standards in data collection, sharing, and use.

Potential Impacts on Various Industries

In the following, I will go through just a few fields of how AI can and has already impacted, or upon its release, has been theorized about possibilities of good and ill effects. But I wanted to include these as many people have said one sentence: “This is going to affect jobs.” But I have done so myself, not gone through some fields and given examples of how it will affect them.?

Healthcare:

Benefits: AI can transform healthcare in numerous ways, with help and potential drawbacks. Here are some of the key ways AI can benefit and damage healthcare:

1. Early diagnosis and disease prediction: AI algorithms can analyze medical images, genomic data, or electronic health records to identify patterns, enabling earlier diagnosis and prognosis of diseases such as cancer, Alzheimer’s, or diabetes.
2. Personalized medicine: By analyzing vast amounts of patient data, AI can help identify the most effective treatments for individual patients, taking into account their genetic makeup, lifestyle, and medical history.
3. Drug discovery: AI can accelerate the process of drug discovery and development by identifying potential drug candidates, predicting their effectiveness, and optimizing their chemical structures.
4. Telemedicine and remote monitoring: AI-powered chatbots, wearable devices, and mobile apps can provide remote healthcare services, monitor patient health, and deliver personalized advice or interventions, improving access to care and reducing healthcare costs.
5. Enhanced medical imaging: AI algorithms can improve the quality, speed, and accuracy of medical imaging by enhancing image resolution, reducing noise, or automating the detection and interpretation of abnormalities.
6. Clinical decision support: AI systems can assist healthcare professionals in diagnosing, treating, and managing diseases by providing real-time insights, recommendations, or alerts based on evidence-based guidelines, best practices, or relevant patient data.
7. Healthcare administration: AI can streamline administrative tasks such as scheduling, billing, or patient data management, reducing the workload on healthcare professionals and freeing up time for patient care.

Potential Damages to Healthcare as a Consequence:

1. Loss of human touch: The increasing reliance on AI and automation in healthcare may lead to a depersonalization of care and a loss of empathy, communication, and trust between patients and healthcare providers.
2. Misdiagnosis and errors: AI systems can make mistakes or fail to recognize rare or complex cases, leading to incorrect diagnoses, inappropriate treatments, or missed opportunities for intervention.

To maximize the benefits and minimize the risks of AI in healthcare, it is essential to promote responsible AI development and use, ensure data quality and representativeness, protect patient privacy and security, and maintain human oversight and professional judgment in healthcare decision-making.

Programming and Cybersecurity

AI will both enhance but also come to replace the need for the number of programmers that are required for programming projects. As well, the length of the project will significantly be reduced, although QA time may increase. Having been a Sr. Software Engineer for ten years, I have created a project that would take two weeks, a total of three minutes. However, I also prototyped a website rebuild that was a three-year-long project. I found I could get it done in about a month, Alone not with a group of 3 programmers, a team leader, a QA lead, a QA tester, a Project Manager, and an Agile Director. The Agile names may be different. Sorry, it has been a couple of years.

Benefits of AI in Computer Programming and Cybersecurity:

1. Enhanced productivity: AI-powered tools can automate repetitive tasks and improve efficiency in computer programming, allowing developers to focus on more complex and creative aspects of their work.
2. Improved code quality: AI can analyze code for potential errors, vulnerabilities, or inefficiencies, helping developers to maintain high-quality code and reduce the likelihood of bugs or security issues.
3. Personalized learning and skill development: AI-based learning platforms can provide tailored resources, exercises, and feedback for individuals seeking to improve their programming and cybersecurity skills.
4. Advanced threat detection: AI algorithms can identify and analyze patterns in large datasets to detect potential cyber threats or attacks more quickly and accurately than traditional methods.
5. Advanced Forensic Methods: using programming and the computer’s intelligence and logging the various systems being attacked, it is possible to gain more advanced and faster forensics.?
6. Proactive security measures: AI can predict and prevent potential security vulnerabilities by identifying patterns and trends in historical data, enabling organizations to take proactive steps to protect their systems and data.
7. Real-time incident response: AI can automate and optimize incident response processes, reducing the time it takes to detect, contain, and remediate cyber threats.
8. Enhanced user authentication: AI can improve the effectiveness of user authentication methods by analyzing biometric data, user behavior, or contextual factors to identify and prevent unauthorized access.

Robotics

AI can significantly benefit robotics by enhancing robotic systems’ capabilities, adaptability, and autonomy, leading to various applications and technological advancements. However, there can also be potential downsides to consider.

Benefits:

1. Advanced perception and object recognition: AI algorithms can process data from sensors, cameras, and other devices to enable robots to perceive and understand their environment, recognize objects, and navigate complex spaces.
2. Improved decision-making and planning: AI can empower robots to make more informed decisions and plans based on their understanding of the environment, the task at hand, and their internal states or constraints.
3. Natural language understanding and communication: AI can enable robots to understand and process natural language, allowing them to interact more effectively with humans and perform tasks that require language comprehension.
4. Learning and adaptation: AI can help robots learn from experience, adapt to new situations, and improve their performance over time through techniques like reinforcement learning, imitation learning, or transfer of knowledge.
5. Human-robot collaboration: AI can facilitate more effective collaboration by allowing robots to understand human intentions, predict human behavior, or adapt to human preferences and needs.
6. Autonomous operation: AI can enhance the autonomy of robotic systems by enabling them to operate without continuous human supervision or intervention, making them suitable for tasks that are hazardous, remote, or tedious for humans.
7. Advanced manipulation and locomotion: AI can enable robots to perform more sophisticated and complex manipulation tasks or to adopt more efficient and versatile locomotion strategies, such as walking, climbing, or flying.

Downsides that affect Robotics:

1. Safety and control: The development of intelligent and autonomous robots raises concerns about their safety, potential misuse, and the ability to maintain control over them, especially in the context of military applications or powerful general-purpose robots.
2. Maximizing benefits while minimizing potential downsides of AI in robotics, promoting responsible AI development and use, ensuring equitable access to technology, addressing ethical and social concerns, and maintaining human oversight and control over robotic systems are essential.

AI for Agricultural Use

AI can bring numerous benefits to agriculture, helping to optimize various aspects of crop and livestock management and promoting sustainable farming practices. However, there can be potential adverse consequences as well.

Benefits:

1. Precision agriculture: AI-powered tools can collect and analyze data from sensors, drones, and satellites to monitor crop health, soil conditions, and weather patterns, allowing farmers to make more informed decisions about planting, irrigation, fertilization, and pest control.
2. Crop yield prediction: AI models can analyze historical and real-time data on climate, soil, and crop performance to predict crop yields and optimize resource allocation, helping farmers to maximize productivity and minimize waste.
3. Pest and disease detection: AI algorithms can identify conditions in agriculture, such as pests, diseases, or nutrient deficiencies in crops. Some methods would be capable of analyzing images captured by drones or ground-based cameras, enabling early intervention and targeted treatment.
4. Livestock monitoring: AI has the potential to monitor the health, behavior, and productivity of livestock through wearable devices, cameras, or automated systems, improving animal welfare and farm management.
5. Autonomous farm equipment: AI-powered tractors, harvesters, or robotic systems can perform tasks such as planting, weeding, or harvesting more efficiently and accurately than traditional methods, reducing labor requirements and operational costs.
6. Supply chain optimization: AI can help optimize agricultural supply chains by predicting demand, reducing spoilage, and improving transportation and logistics, leading to more efficient and sustainable food systems.
7. Environmental sustainability: AI can support sustainable farming practices by optimizing resource use, reducing waste, and minimizing the environmental impact of agriculture, for example, by promoting precision irrigation or targeted pesticide application.

Adverse Consequences of all the topics mentioned above coincide with the fears of the technological experts warning of the advancement beyond GPT-4 and Bard:

1. Job displacement: The increased use of AI and automation in agriculture may displace jobs or change the nature of work for farmers and farm workers, potentially leading to unemployment or the need for re-skilling.
2. Digital divide: Access to AI technologies may be limited by factors such as infrastructure, affordability, or digital literacy, potentially exacerbating existing inequalities and disparities in agricultural productivity and income.
3. Data privacy and security: The collection and use of farm data by AI systems may raise concerns about data privacy, ownership, and security, as well as the potential for data misuse or commercial exploitation.
4. Ethical and social concerns: The use of AI in robotics raises moral and social situations, such as the potential for biased decision-making, the erosion of privacy, or the impact on human dignity and social relationships.
5. Dependence on technology: Overreliance on AI and digital technologies may make farmers more vulnerable to technical failures, cyberattacks, or loss of traditional farming knowledge and skills.
6. Environmental impact: While AI can contribute to more sustainable agriculture, it may also enable more intensive farming practices that could have negative ecological consequences, such as increased pesticide use, water consumption, or land conversion.
7. Privacy concerns: The use of AI in cybersecurity often requires the collection, analysis, and storage of large amounts of sensitive data, raising concerns about privacy and data protection.
8. Accountability and transparency: As AI systems become more complex and autonomous, holding them and their creators accountable for their actions and decisions becomes increasingly challenging, especially in cybersecurity.
9. Adversarial attacks: Cybercriminals may exploit the vulnerabilities of AI systems or use malicious machine learning techniques to deceive or compromise AI-driven security measures.

We must promote responsible AI development and use and ensure that technology access is balanced between all classes, people, and countries to allow intelligence to be effectively ‘taught fully.’

References

• Bias and discrimination in AI:
• a. Joy Buolamwini - https://www.media.mit.edu/people/joyab/overview/
• b. Timnit Gebru - https://ai.stanford.edu/~tgebru/
• Job displacement and labor market disruption:
• a. McKinsey Global Institute - www.mckinsey.com/mgi
• b. World Economic Forum - www.weforum.org
• Privacy and surveillance:
• a. Electronic Frontier Foundation (EFF) - https://www.eff.org
• b. Center for Internet and Society at Stanford Law School - https://cyberlaw.stanford.edu
• Accountability and transparency in AI:
• a. AI Now Institute - https://ainowinstitute.org
• b. OpenAI - https://www.openai.com
• AI safety and control:
• a. Future of Life Institute - futureoflife.org
• b. Partnership on AI - https://www.partnershiponai.org
• Ethical considerations in AI:
• a. AI Ethics Lab - https://aiethicslab.com
• b. IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems - https://standards.ieee.org/industry-connections/ec/autonomous-systems.html
• Environmental impact of AI:
• a. Google AI - https://ai.google/research/areas/environment/
• B. Climate Change AI - https://climatechange.ai
• Data pollution in AI:
• a. IBM Research - https://www.research.ibm.com/artificial-intelligence/data-quality/
• b. Microsoft AI - https://www.microsoft.com/en-us/research/research-area/ai/data-quality/
• AI in healthcare:
• a. DeepMind Health - https://deepmind.com/applied/deepmind-health/
• b. IBM Watson Health - https://www.ibm.com/watson-health
• Disagree (regarding potential damages):
• a. American Medical Association - https://www.ama-assn.org
• b. World Health Organization - https://www.who.int
• AI in programming/cybersecurity:
• a. Darktrace - https://www.darktrace.com
• b. Palo Alto Networks - https://www.paloaltonetworks.com
• AI in robotics:
• a. Boston Dynamics - https://www.bostondynamics.com
• b. Softbank Robotics - https://www.softbankrobotics.com
• AI in agriculture:
• a. Blue River Technology - https://www.bluervt.com
• b. The Climate Corporation - https://www.climate.com