A Future with Artifical Intelligence

Artificial intelligence (AI) has existed conceptually since the time of antiquity (McCorduck 4-5) evidenced by descriptions of Talos, an ancient humanoid automaton used to protect the shores of Europa. You will perhaps remember the portrayal of Talos from the movie Jason and the Argonauts - when all of the men are running scared on the beach. AI of course wasn’t formally described theoretically until Alan Turing published his seminal paper “Computing Machinery And Intelligence” in 1950, where the famous ‘Turing Test’ was first introduced. The idea of using AI to solve our problems has been around for a long time.

We have come extraordinarily far in the field of AI since the 50s - so far that we almost don’t recognize it anymore. In antiquity, AI was an imaginary mechanical owl or humanoid. Later it became a pretend robot helping a little boy named Will Robinson or one helping the Jetsons. Over the years, AI has been depicted in countless ways. Sometimes helpful, sometimes horrifying, but always imaginary. It became The Iron Giant, The Terminator, the robot from I,Robot, it became The Matrix. And then out of nowhere, it became real - but not like anything we recognized.

The emergence of, lets call it ‘practical AI has come in the form of web apps like ChatGPT, computer vision systems that enable Tesla autopilot, image editing tools used in film and photo editing software like that provided by adobe, and art generation tools such as Dall-E. Perhaps you’ve seen an advertisement or heard about these technologies on the news. AI has also been applied in countless places that normal people perhaps wouldn’t expect. Farming equipment, customer support software, building air conditioning systems, your phone’s touch screen - and all manner of other random places you might not expect. The reality is that AI is actually already everywhere in this modern digital age. We are all just becoming more and more aware, which has revitalized the imaginations of everyday people for better and for worse.

Like most (if not all) technology, Artificial Intelligence is a double edged sword. If wielded responsibly, it can be an unbelievably useful tool that defines a new phase in the general technological advancement of the human species. If wielded irresponsibly or in malice, it could destroy us all. And like any technology, it is impossible to entirely regulate its use across all borders - there is no way to fully prevent the abuse of AI by hostile individuals or even nation states. And despite how incredible some of these modern tools may seem - we are currently still in the absolute infancy of AI - like a baby who has just grown from a helpless blob to one that can begin to track an object held before their eyes.

And yet, an enormous number of people have found great benefit from using tools such as ChatGPT. This tool in particular has proven itself to be an incredible productivity tool when used to generate ideas, provide explanations, and even collaborate with. I myself recently found great success in using ChatGPT when authoring the C# computer programming language library Sailfish - a tool that is now available to and used by companies to test how quickly their applications run. Popular productivity tools such as Whimsical.com use AI to help generate ideas and mind maps, Microsoft has integrated AI into its core search product offering Bing. AI has found applications in industry, the military, art, and just about anywhere else you can imagine.

The advantage that AI gives us is that it has the potential to outperform us in nearly every way. Some of these ways will be obvious. AI may give us faster, more reliable medical diagnostics or give us a faster, more intuitive human-like experience when searching for information. But artificial intelligence may also be the key to unlocking faster scientific progress in ways that are not obvious. AI can distill gaps in our knowledge across entire domains - tens of thousands of recent scientific publications - and inform us where we need to investigate further faster than any PhD student ever could. AI can likely form new insightful conclusions based on existing data, finding information we may already be overlooking. AI can literally invent new technologies for us to manage global issues like carbon-capture to prevent global warming. We have already begun using AI to optimize control in historically unstable, and therefore unusable, fusion reaction experiments. AI can help us design new medical compounds by seeing things that we have either overlooked, or by performing data analysis in a way that is impossibly complex for a human.

The reason for this is because AI is a technology that allows computers to do what humans do - approximate things. AI is built using digital Neural Networks - and neural networks satisfy the Universal Approximation Theorem - in the same way the human brain does. This is what gives AI the potential to learn like a human, and create outputs like a human.

It is perhaps this human element to AI that has caused it to also become the source of fear and anxiety for those who are uncertain about how AI will evolve. We know that it may be used for great good. (See Appendix 1 for an AI-generated non-exhaustive list of ways that AI can or currently does benefit humanity) But we also know it could be used for great evil. And while we’re not anywhere near the level of, say, The Terminator - there are a large and growing number of obvious and insidious ways that AI could be used to harm others. (See Appendix 2 for an AI-generated non-exhaustive list of ways AI can hurt us)

One popular way that AI has been abused is through the creation of unsolicited deep fakes, where someone takes a video of one person and uses it to superimpose that person’s appearance on the face or body of someone in another video. This has been demonstrated to be used effectively in a range of scenarios both good and bad - from fun face swaps or putting yourself into your favorite movie… to maliciously putting someone in a pornographic video against their will or creating false but incredibly convincing impersonation videos of global leaders. You can easily imagine how deep fakes could be abused to destroy the reputation of someone, or even create mass confusion for millions of people.

There are perhaps even more insidious ways that AI can negatively impact the world. One major fear - that has in fact become a reality for some already - is the wholesale replacement of workers by AI. This is a complex problem that has immediate impacts on both the individual as well as far reaching consequences for entire economies. What happens when AI becomes sufficiently good to replace large portions of the workforce? For example - one company recently eliminated 90% of its support staff, which was replaced by a service such as ChatGPT. A great success for the company, but 23 people left unemployed in a climate where all of the other companies are looking to eliminate the same roles that they might otherwise apply for. This may lead to a situation of mass unemployment.

So what is being done about this incredible, and perhaps incredibly dangerous technology? Is it being regulated? Is it being controlled? What kind of protections do the normal people have against such a looming potential economic catastrophe?

AI is already regulated to various degrees implicitly, but the US government does not (yet) have any explicit laws regulating the application of AI. For example, deep fakes can be made by anyone who knows how to use the internet - and yet their more heinous applications are completely unregulated. Fortunately, we have seen regulators limit the pace at which things like autonomous vehicles have been adopted and some existing laws can be interpreted by the judiciary with AI in mind. Furthermore, there are good reasons to believe that more deliberate regulation is coming. For example, we saw recent testimony from the CEO of ChatGPTs owner company OpenAI who himself made the case for the regulation of AI.

So where does that leave us, the people?

It is critical to recognize and accept that AI is not going away. It is too valuable as a tool for good, and it is too effective as a weapon of evil. We’ll all also need to accept sooner or later that most of the things we think are possible with AI - both exciting and freighting - are eventually going to be possible. We may never have the futuristic materials used to create the robots of I,Robot, but we’ll certainly be capable of producing robotic automatons like Talos that exhibit human-like capabilities. Indeed, some are arguing that existing generative tools such as ChatGPT already demonstrate that human-like spark.

The only way for humanity to navigate this new, unbelievable, world alongside AI is for us all to first try and understand it. AI, at its most fundamental, is nothing more than a marriage between computer science and mathematics. When applied, it is deployed through applications which are also built with programming languages. When it interacts with the real world, it does this through machines like robotics. So for the people to understand AI in order to regulate it and protect ourselves from it, it is absolutely necessary that we develop a general level of literacy in these domains. That is a difficult problem in and of itself. If our legislators are going to be effective at understanding how, why, and when to regulate AI, then they are going to need to understand what is possible with AI and more or less how it works as well. Ignorance is not an option.

Ultimately, the coexistence of AI and humanity necessitates a balance between harnessing its power for good while mitigating its potential for harm. It is crucial for society to recognize and accept that AI is here to stay. Instead of fearing its advancement, we must actively work towards responsible development, ethical deployment, and effective regulations. By doing so, we can ensure that AI continues to be a transformative tool for progress and avoids becoming a destructive force. The path forward lies in embracing AI's potential while addressing its challenges, fostering collaboration between researchers, policymakers, and society at large to shape a future where AI serves the greater good while protecting human interests.

By the way - did you realize that last paragraph was written entirely by AI?

Additional Work Cited

McCorduck, Pamela. Machines Who Think: A Personal Inquiry Into the History and Prospects of Artificial Intelligence. Taylor & Francis, 2004.

Appendix 1

The following are two prompts with their responses. Taken from https://chat.openai.com/

Prompt

List all the ways you can think of that AI can help humanity

ChatGPT Response

AI has the potential to benefit humanity in numerous ways across various domains. Here are some of the ways AI can assist:

1. Healthcare: AI can help in diagnosing diseases and identifying patterns in medical images, enabling early detection of illnesses. It can also assist in drug discovery, personalized medicine, and patient monitoring.

2. Education: AI can provide personalized learning experiences, adaptive tutoring, and intelligent feedback to students. It can also assist teachers in automating administrative tasks and developing tailored curricula.

3. Transportation: AI can improve transportation systems by optimizing traffic flow, predicting maintenance needs, and enhancing autonomous vehicle technology for safer and more efficient journeys.

4. Environment: AI can aid in monitoring and analyzing large-scale environmental data, such as climate patterns and deforestation, to provide valuable insights for conservation efforts and sustainable development.

5. Agriculture: AI can optimize crop management, predict weather patterns for better farming decisions, and automate tasks like irrigation and harvesting. It can also assist in disease detection in plants and livestock.

6. Finance: AI can improve fraud detection, risk assessment, and algorithmic trading. It can also offer personalized financial advice and streamline customer service through chatbots.

7. Manufacturing: AI can enhance automation and robotics in manufacturing processes, leading to increased productivity, quality control, and worker safety. It can also optimize supply chain management and predictive maintenance.

8. Cybersecurity: AI can help identify and respond to cyber threats by analyzing patterns, detecting anomalies, and enhancing network security. It can also assist in fraud detection and data protection.

9. Natural language processing: AI-powered language models can enable better communication between humans and machines, facilitating tasks like language translation, voice assistants, and sentiment analysis.

10. Scientific research: AI can assist in analyzing vast amounts of scientific data, accelerating discoveries in fields such as genomics, astronomy, and material science. It can also simulate complex systems and aid in drug development.

11. Social services: AI can be used in social welfare programs to analyze data and identify individuals or communities in need. It can help optimize resource allocation, improve service delivery, and identify potential risks.

12. Personal productivity: AI-powered virtual assistants can help with organizing schedules, managing tasks, and providing relevant information, thereby increasing individual productivity and efficiency.

13. Entertainment: AI can enhance gaming experiences, create realistic virtual worlds, and assist in content curation and recommendation systems for music, movies, and other media.

14. Disaster response: AI can support disaster management by analyzing real-time data, predicting and monitoring natural disasters, and coordinating relief efforts efficiently.

These are just a few examples of how AI can positively impact humanity. The potential applications of AI are vast and continue to expand as technology advances and new possibilities emerge.

Prompt

List the ways that AI is currently being used

ChatGPT Response

AI is already being employed in numerous real-world applications across various industries. Here are some of the ways AI is currently being used:

1. Virtual assistants: AI-powered virtual assistants like Siri, Google Assistant, and Alexa use natural language processing and machine learning algorithms to understand and respond to user queries and perform tasks.

2. Image and speech recognition: AI is used extensively in image and speech recognition applications. Examples include facial recognition systems for security, voice assistants for speech-to-text conversion, and object recognition in self-driving cars.

3. Recommendation systems: AI algorithms are employed in recommendation systems used by platforms like Netflix, Amazon, and Spotify to provide personalized content suggestions based on user preferences and behavior patterns.

4. Fraud detection: AI is used in fraud detection systems in finance and e-commerce sectors to identify suspicious transactions and activities by analyzing patterns and anomalies in data.

5. Autonomous vehicles: AI plays a crucial role in self-driving cars and autonomous vehicles. It enables object detection, path planning, and decision-making processes necessary for safe navigation on the roads.

6. Natural language processing: AI is used in natural language processing applications such as language translation, sentiment analysis, and chatbots, enabling more effective human-computer interaction.

7. Healthcare diagnostics: AI is utilized in medical imaging for tasks like identifying abnormalities in X-rays, CT scans, and MRIs. It can also assist in diagnosing diseases by analyzing patient data and symptoms.

8. Customer service chatbots: AI-powered chatbots are used by companies to handle customer inquiries, provide support, and automate customer service processes, improving response time and efficiency.

9. Financial trading: AI algorithms are used in algorithmic trading to analyze market data, predict trends, and make high-speed trading decisions.

10. Social media analysis: AI is employed to analyze social media data, detect trends, and perform sentiment analysis to understand public opinion and support marketing strategies.

11. Personalized advertising: AI is used in digital advertising to analyze user behavior and preferences, enabling personalized ad targeting and content recommendations.

12. Language translation: AI-powered language translation systems like Google Translate utilize machine learning algorithms to provide automated translations between different languages.

13. Cybersecurity: AI is used in cybersecurity applications to detect and prevent cyber threats. It can analyze network traffic, identify patterns of suspicious behavior, and respond to potential security breaches.

14. Manufacturing and robotics: AI is employed in manufacturing processes to optimize production lines, automate tasks, and improve quality control. Robotics also utilizes AI for object recognition and decision-making.

These are just a few examples of how AI is currently being used in various sectors. The applications of AI continue to grow and evolve as technology advances and new opportunities arise.

Appendix 2

Prompt

What are the ways that AI could be used maliciously?

ChatGPT Response:

While AI has the potential for significant positive impact, it can also be misused for malicious purposes. Here are some ways AI could be used maliciously:

1. Deepfakes: AI can generate highly realistic fake videos, images, and audio, which can be exploited for spreading misinformation, fake news, or defaming individuals.

2. Cyberattacks: AI can be used to develop more sophisticated and targeted cyberattacks. AI-powered malware can adapt and evolve, making it challenging to detect and defend against.

3. Social engineering: AI can be employed to create more convincing phishing emails, scam calls, or messages, increasing the effectiveness of social engineering attacks and fraud.

4. Autonomous weapons: AI can be integrated into autonomous weapons systems, which could be programmed to identify and attack targets without human intervention, leading to ethical and humanitarian concerns.

5. Privacy invasion: AI can be used to analyze large amounts of personal data without consent, potentially compromising privacy and violating individuals' rights.

6. Bias and discrimination: If trained on biased or discriminatory data, AI algorithms can perpetuate and amplify existing biases, leading to discriminatory outcomes in areas such as hiring, lending, and law enforcement.

7. Manipulation of financial markets: AI algorithms can be used to manipulate stock prices, engage in high-frequency trading for unfair advantages, or generate misleading market predictions.

8. Autonomous hacking: AI algorithms can be used to automate the process of finding and exploiting vulnerabilities in computer systems, accelerating the rate at which cyberattacks occur.

9. Manipulation of public opinion: AI algorithms can be used to manipulate social media feeds, spreading disinformation, propaganda, or engaging in astroturfing to shape public opinion.

10. Surveillance and tracking: AI-powered surveillance systems can be used to monitor and track individuals without their knowledge or consent, leading to privacy infringements and potential abuses.

It is crucial to recognize these potential risks and work towards developing ethical frameworks, regulations, and safeguards to mitigate the malicious use of AI and ensure responsible deployment of these technologies.