INTRODUCTION                                                                                                                              

AI has become increasingly relevant across many fields, with recent innovations like ChatGPT by OpenAI in 2020. This increased integration will have many implications going forward, including what we do on a day-to-day basis, what we need to know and do as engineers, and how we should guide this integration as technology progresses. This prevalence has also resulted in many concerns and misconceptions from various people regarding AI. Hopefully, this workshop can shed some light on what AI is and what it can be used for, as AI will likely only become increasingly important in the future.

LEARNING OBJECTIVES                                                                                                                 

After this workshop, you should be able to:

1. Recognize key points in the development of AI technology
2. Understand the key components of AI
3. Describe the advantages and disadvantages of AI technology
4. Discuss the role of Engineers in the development of AI technology

HISTORY OF AI AND AI TECH                                                                                                       

Technology Timeline

• Artificial Neural Networks (ANN) – These are at the base of AI concepts as they are simply a way of looking at data inspired by how the human brain works. When humans learn things, we make new connections within our brains, linking the new information together with old information. These artificial neural networks aim to do the same thing.
• Deep Learning – This is a step up from ANNs, using these neural network links for pattern recognition purposes. The key feature of these deep learning programs is their use of multiple layers of these neural networks, allowing for very complex connections to be formed.
• Machine Learning – Machine Learning is similar to Deep Learning, but is often less complex, relying on a human to sort the “relevant” data. They often deal more with statistics, and thus are used in places such as “predictive analytics”.
• Artificial Intelligence – AI is the culmination of all of these ideas, as an attempt to recreate/simulate human intelligence.

AI Development Summary

So as you can see, humans have always used computers to solve problems. The crux of the matter is that we want computers to be able to solve more and more complex problems in increasingly efficient ways.

Most programs you interact with are likely just algorithms, or simply steps that a computer follows to complete a task. For things like AI, in an effort to have computers solve problems in a more human-like fashion, we’ve tried to program them to approach these challenges more like humans would. This led to the invention of Artificial Neural Networks, Deep Learning and Machine Learning. And these processes of thinking are components of artificial intelligence. Nowadays, machine learning and deep learning have paved the way for more complex versions of “Generative AI”, that is, AI capable of generating new content. Historically, this might have been simple chatbots, but now, it can create human-like text, images, video, and even audio.

WHAT CONSTITUTES ARTIFICIAL INTELLIGENCE?                                                                

At its core, AI is just simulated human intelligence. This, however, comes with its own caveats as “human” intelligence is both very vague and very broad. With the goal being intelligence as humans would use, there is a lot of focus on cognitive skills like the ability to learn, the ability to reason, the ability to self-correct, and the ability to be creative. So going back to the definition, an AI should be able to do all of these things, and do them in a way in which humans would

The term artificial intelligence, however, is often used quite differently, mostly often referring to a select few components of AI. For instance, many stand-alone applications of machine learning or neural networks are being labelled as “AI”, while being very limited in scope. AI is quite a buzzword in media nowadays, and we’ve seen similar things in the past. The term “smart” device previously occupied a similar space in the tech industry, with the label “smart” being applied to anything and everything capable of sensing something. Likewise, AI is being applied to anything that uses components of AI, so it can be helpful to keep this in mind when hearing about these kinds of developments.

CLASSIFYING AI                                                                                                                              

There currently exist three main categories for classifying AI. Like the Venn diagram you saw before for neural networks, deep learning and machine learning, AI can be classified as artificial narrow intelligence, artificial general intelligence, and artificial super intelligence.

• Narrow Intelligence (Weak AI) – These types of AI are trained to do a specific task with human-like capabilities. This describes all the AI that we have developed to date. Even something like ChatGPT is only designed to take user input and give you a human-like text response.
• Artificial general intelligence (Strong AI) – These types of AI possess the ability to function completely like a person would, thus learning on their own from unspecific data sources, generalizing information from one task to another. Currently only exists in fiction (Jarvis, Wall-E).
• Artificial Super Intelligence – Having our cognition abilities, while having better memory, data processing and analytical capabilities allow this category of AI to be far superior to us in problem-solving. This category, just like the last, is purely fictitious for now but may have some serious implications for the future.

So, as we currently live in a world of rapidly developing weak AIs, with strong AI being a bit out of our reach currently, the question becomes “What can we do with these technologies?”.

_{Naveen Joshi, “7 Types of Artificial Intelligence”. Forbes. Available from: https://www.forbes.com/sites/cognitiveworld/2019/06/19/7-types-of-artificial-intelligence/?sh=239af750233e [Accessed 5 Feb, 2024].}

Benefits of AI Technology

Don’t be afraid! Media tends to sensationalize AI, and a lot of media tends to tell you why you should be scared of it. While there are some disadvantages for sure, there are also a lot of potential benefits.

As it stands, AI is great for replacing humans in Dull, Dirty, and Dangerous jobs. Humans are very susceptible to fatigue and injury, and thus AI is a great fit for these types of jobs, as they don’t suffer from the same types of problems. Additionally, instead of replacing jobs, AI has the ability to enhance existing ones.

Think, for instance, about a field like medicine. Doctors currently do great work in both diagnostics and talking patients through procedures or conditions. But doctors are people too, and their time is often limited. Having access to an AI tool that can analyze symptoms and give likely causes would be a great pre-screening tool, or even provide greater access to healthcare for those who need it. As a computer doesn’t forget things as a human would, this would give doctors more options to think about that they might not have considered, hopefully increasing their success in making correct diagnoses.

AI as a tool would also enable us to tackle problems we couldn’t have tackled before. There have already been great strides in protein folding and cancer research thanks to machine learning, and with AI becoming increasingly more refined by the day, tools for climate simulation, or even machine vision, are bound to only get better, helping us tackle challenging problems much more easily.

_{Henny Admoni, “Robotics Professor Answers Robot Questions from Twitter”. WIRED. Available from: https://www.youtube.com/watch?v=r4rHAqsqF80&ab_channel=WIRED [Accessed 5 Feb, 2024]}

Disadvantages of AI Technology

Don’t be entirely afraid! All of this being said, there are still drawbacks to AI. Even though they’re not immediately coming for all of our jobs, there will definitely be an effect on some job markets, especially lower-skilled job markets on which many people rely.

There is also a huge issue with something that we discussed previously. As AI advances and neural networks become more and more complicated containing more and more layers, we become unable to know what kind of connections they’re making. This leads to issues, particularly when setting up safeguards against harmful information. This has been the cause of an entirely new field of getting information out of AI, called prompt engineering. Through prompt engineering, people can bypass security restrictions on data provided by AI, thereby obtaining potentially harmful information.

People can also use this information for malicious actions, in addition to misusing AI to worsen existing problems, like phishing, deepfakes, and copyright infringement. On top of all of this, AI can sometimes just be wrong, or even lie to users. Because we don’t know what connections they are making, it’s very easy for them to fabricate false information, or distribute incorrect information due to issues with the data they’re being trained on. This has the potential to make AI very unreliable.

ACTIVITY 1: AI Advantages and Disadvantages Case Study                                                   

Based on one of the case studies given on the provided worksheet, discuss the advantages/disadvantages of using AI in this field. Examples of situations in which you could use AI in each of these fields are given but feel free to argue for your own scenario. Consider what category of AI you would need to accomplish this task and the challenges that may be involved, in order to better answer the discussion questions.

• Activity 1 Worksheet: [Link to Worksheet]

Activity 1 – Key Takeaways

• AI has many advantages and disadvantages across many different fields
• AI is not always the most suitable, in its current form of Weak AI
• There are many things we need to mitigate against going forward as AI develops

THE ROLE OF ENGINEERS                                                                                                           

Impact of AI on Engineering

We’ve seen that AI can be a very useful tool for a variety of disciplines, and can be very dangerous if not properly managed. AI also has a few implications for the field of engineering, as to how it affects our careers, our practical tasks, and how we go about ethics.

• Career implications – What engineers are required to know, and to do. This is inclusive of how we educate future engineers, and new fields that might develop within the sphere of AI technology, such as the field of prompt engineering that we discussed previously.
• Practical implications – How AI gets integrated into our daily lives. In the future, we might be using AI for anything, ranging from our morning routines, to getting news, or even getting work done. As engineers, we have the responsibility of properly ushering in this new technology while defending against misuse. This is especially crucial as AI becomes more and more intertwined with our activities, as these AI will have access to a lot of our information.
• Ethical Implications – When developing things, we should always consider if they should exist in the first place, and if they should, how we can properly implement them to prevent accidents or misuse. For things like security, maybe AI capable of storing a lot of user data is unnecessary, even if it’s convenient.

TECHNOLOGICAL STEWARDSHIP                                                                                              

While ethics is a very complicated subject, there are a few technological stewardship principles that we can apply to maximize the benefit of technologies like AI for everyone. Quite fittingly, these 4 principles are Tech STEW.

• Seeking purpose allows us to consider how the use of the technology resonates with you as the developer, we as a team, and us as a community. In other words, it allows us to truly understand the purpose of developing it.
• Taking responsibility involves thinking about how the technology will solve problems, the intended consequences of it, and potential non-intended impacts. Responsible development and implementation involve a lot of critical thinking and foresight, while also requiring plans for mitigating damage when they inevitably occur.
• Expanding inclusion is important in both development and implementation, as diverse teams are better able to think of wholistic solutions and plans, and these solutions must be equitable for anyone who needs this technology. With something like AI, issues with inclusion might come in the form of unequal benefits to those who have easy access to computers, vs those who don’t.
• Widening our approaches allows us to consider many alternatives when developing solutions to problems. This involves not just the methods of problem-solving, but also the means of problem-solving. Considering many options, with many different approaches allows us to settle on the best, most inclusive and responsible solution, with plans for ensuring the safety and wellbeing of society.

By applying these ways of thinking and doing, we as engineers can strive towards better solutions, while constantly trying to improve. As new, unforeseen events occur, these principles can ensure that we can always continue to adapt and grow.

Important Next Steps

With these stew principles in mind, we can consider the next steps we should be taking as engineers, in the development and implementation of new AI technologies. Thinking back to the tech timeline, there has been exponential growth in AI developments as time goes on. With rapid development inevitable, it becomes very important to lay down better regulations, especially when it comes to their use and transparency.

As a user, we should be privy to the information that these technologies can collect from us, and how they are being used and handled. This kind of transparency is crucial for maintaining user security.

Additionally, while strides have been made by the government of Canada to make policies outlining when and where AI can be used, larger organizations and countries need to come together to formalize many of these regulations. Accountability is also crucial in this process, as these policies won’t help if they are never strictly enforced.

And finally, we need to accept that AI is likely here to stay. As discussed in the section regarding the benefits of AI, AI isn’t all doom and gloom as it can be a very useful tool if used correctly. Thus, we should likely start training people on AI, and when/how to use it. If it’s going to continue to grow regardless of what we do, we might as well take full advantage of it in a responsible manner that can help advance society.

ACTIVITY 2: Modern AI Use and its Implications                                                                      

In your groups, discuss some examples of the use of AI in modern times, along with their implications. Try to consider scenarios in which AI was used, and it had either a positive or negative effect on a situation. Relevant AIs which you might be familiar with include ChatGPT, Stockfish/Deep Blue, or even AI personalities/streamers.

• Activity 2 Worksheet – [Link to Worksheet]

Activity 2 Key Takeaways

• AI is being used in a wide variety of applications even though they are currently limited in scope
• These uses often have many unintended consequences in addition to their intended consequences
• Laws and regulations are important, especially regarding user privacy and safety

WORKSHOP SUMMARY                                                                                                               

Just to wrap things up, today we’ve learned a bit about the history of AI technology, along with what constitutes AI and common misconceptions about AI. AI is likely here to stay and has many potential benefits, but an equal amount of potential consequences we need to mitigate against. As such, we as engineers must be mindful of tech stewardship principles, ensuring that we continue to develop these technologies in the right direction, for the greatest benefit to all of society.

RESOURCES                                                                                                                                    

Click here to access the workshop worksheets.

CONGRATULATIONS! YOU HAVE REACHED THE END OF THIS WORKSHOP!