The talk which this webpage supports and on which it is based was given by Geoff Sheath to the Get Together Monday group at Shrewsbury House Community Centre, Greenwich, SE18 3EG on Monday 10 July 2023.
Created in 1770, this chess-playing automaton toured Europe and beyond beating most of the best chess players. Only after it had been destroyed 80 years later was it revealed that there was a chess master hidden inside.
Charles Babbage with help from Ada Lovelace drew up plans for the Analytical Engine between 1840 and 1870. Although mechanical it was arguably the first ever programmable computer. Due to shortage of funds and construction difficulties it was never completed by Babbage but the Science Museum made a copy many years later.
Alan Turing in 1950 proposed the following test to decide if a machine was intelligent.
A human evaluator would judge the conversations between a human and a machine
The evaluator would know that it was a machine talking with a person
The conversation would be text-only
If the evaluator could not reliably tell the machine from the human, the machine would be said to have passed the test.
The test results would not depend on the machine's ability to give correct answers to questions, only on how closely its answers resembled those a human would give.
adapted from Wikipedia
This was a computer program that was designed to emulate a particular style of therapy. Try it and see if you think it passes the Turing test.
Intelligence demonstrated by computers, as opposed to human or animal intelligence.
"Intelligence" covers the ability:
to learn
to reason
to generalize
to infer meaning
Advanced web search engines (e.g. Google Search),
Recommendation systems (used by YouTube, Amazon, and Netflix)
Understanding human speech (such as Siri and Alexa)
Self-driving cars (e.g. Tesla, Waymo)
Generative or creative tools (ChatGPT and AI art - DALL-E)
Automated decision-making (cancer scans)
Competing at the highest level in strategic game systems (such as chess and Go).
Artificial intelligence was founded as an academic discipline in 1956
It has experienced waves of optimism, followed by disappointment, followed by new approaches, success, and renewed funding.
AI research has tried and discarded many different approaches, including simulating the brain, modeling human problem solving, formal logic, large databases of knowledge, and imitating animal behaviour.
In the first decades of the 21st century, highly mathematical and statistical machine learning has dominated the field, and this technique has proved highly successful, helping to solve many challenging problems throughout industry and academia.
You can register and use ChatGPT free. Go to https://openai.com/chatgpt to sign up and use it in a web browser or there are apps for iPhones and android phones.
Some of the research for this talk were prepared by ChatGPT.
Chat GPT stands for Chat Generative Pre-Trained Transformer and was developed by an AI research company, Open AI
DALL-E-2 is an AI system that can create realistic images and art from a description in natural language.
You can sign up to try it at https://openai.com/dall-e-2 It is no longer free. At the moment it costs $18 for 115 tokens. You use 1 token per picture. It won't allow you to ask for pictures including most celebrities nor offensive images e.g. it refused to create "A photo of Boris Johnson riding a large dog” but you don't use a token if your request is refused!
This photo was created by typing, "A photo of a man riding a large dog” as an instruction to DALL-E.
Deep learning is a subfield of artificial intelligence that focuses on training neural networks to learn and make predictions from vast amounts of data.
It is inspired by the structure and function of the human brain, with multiple layers of interconnected nodes called neurons.
These networks can automatically extract meaningful features from raw data and perform complex tasks, such as image recognition, natural language processing, and speech synthesis.
Deep learning has achieved remarkable breakthroughs in various domains, such as computer vision and language understanding.
By leveraging its ability to uncover intricate patterns and relationships within data, deep learning holds the potential to revolutionise industries and drive advancements in areas like healthcare, autonomous systems, and scientific research.
The two lists below were produced by ChatGPT. These are the full versions: I just copied the headings for my talk.
Healthcare: AI can significantly improve healthcare by enhancing diagnosis accuracy, predicting diseases, and assisting in drug discovery and development. It can analyze vast amounts of medical data to identify patterns and provide personalized treatment recommendations, leading to improved patient outcomes and more efficient healthcare delivery.
Education: AI has the potential to revolutionize education by personalizing learning experiences for students. Intelligent tutoring systems can adapt to individual needs, provide tailored feedback, and support teachers in delivering more effective instruction. AI can also automate administrative tasks, freeing up educators' time to focus on student engagement and creativity.
Environmental Sustainability: AI can contribute to environmental sustainability by optimizing energy consumption, predicting and mitigating natural disasters, and facilitating smart resource management. It can assist in designing energy-efficient buildings, optimizing transportation routes, and monitoring ecosystems to enable more effective conservation efforts.
Accessibility: AI can help create a more inclusive society by providing solutions for people with disabilities. Natural language processing and computer vision technologies can improve communication and accessibility for individuals with hearing or visual impairments. AI-powered assistive technologies can enhance mobility and independence for people with physical disabilities.
Transportation: AI can enhance transportation systems by enabling autonomous vehicles, reducing traffic congestion, and improving safety. Self-driving cars can reduce accidents caused by human error, optimize traffic flow, and enhance fuel efficiency. AI can also assist in managing public transportation systems, improving route planning and predicting maintenance needs.
Agriculture: AI can revolutionize agriculture by optimizing crop yields, reducing resource waste, and improving pest and disease management. By analyzing data from sensors, satellites, and drones, AI can provide real-time insights on soil conditions, crop health, and irrigation needs. This helps farmers make informed decisions and maximize their productivity while minimizing environmental impact.
Cybersecurity: AI can bolster cybersecurity efforts by detecting and mitigating threats in real-time. Machine learning algorithms can analyze network traffic patterns, identify anomalies, and respond to cyberattacks more rapidly than human analysts. AI can also assist in fraud detection, protecting individuals and organizations from financial losses.
Customer Service: AI-powered chatbots and virtual assistants can improve customer service experiences. These intelligent systems can provide instant responses to customer inquiries, offer personalized recommendations, and handle routine tasks, freeing up human agents to focus on more complex issues. This enhances customer satisfaction and efficiency in various industries.
Personalized Marketing: AI can enhance marketing strategies by analyzing vast amounts of consumer data and providing personalized recommendations. By understanding individual preferences and behavior patterns, AI algorithms can deliver targeted advertisements, optimize pricing strategies, and improve customer engagement, leading to more effective marketing campaigns.
Scientific Research: AI can accelerate scientific research by analyzing large datasets, simulating complex systems, and assisting in data interpretation. It can help identify new drug candidates, optimize experimental designs, and predict outcomes in various scientific fields, from astronomy to genomics. AI's ability to discover patterns and generate hypotheses can drive breakthroughs and accelerate scientific progress.
It is important to note that while AI has significant potential to improve society, careful consideration must be given to ethical and privacy concerns, ensuring transparency, accountability, and equitable access to these technologies.
“Write 10 ways that AI might harm society”
Job Displacement: AI has the potential to automate tasks currently performed by humans, leading to job losses and unemployment in various industries. This could result in socioeconomic inequality and the need for widespread reskilling and job transition programs.
Bias and Discrimination: AI systems can inherit biases present in the data they are trained on, leading to discriminatory outcomes. If not properly addressed, AI can perpetuate existing societal biases related to race, gender, or socioeconomic status, potentially reinforcing inequality and unfair treatment.
Privacy Concerns: AI relies on vast amounts of data, and the collection and analysis of personal information raise concerns about privacy. Improper use of AI can result in unauthorized access to sensitive data, surveillance, and potential breaches of privacy rights.
Security Threats: AI can be exploited by malicious actors for nefarious purposes. Adversarial attacks can manipulate AI systems, leading to misinformation, identity theft, or compromised security systems. AI-powered cyberattacks could become more sophisticated and challenging to detect and defend against.
Dependence on AI: Overreliance on AI systems without proper fail-safe mechanisms can leave society vulnerable to catastrophic failures. In critical domains such as healthcare or transportation, relying solely on AI decision-making without human oversight could have dire consequences if the technology malfunctions or makes erroneous judgments.
Ethical Dilemmas: AI can present complex ethical dilemmas, particularly in areas like autonomous weapons systems. The development and use of AI-powered weapons raise concerns about accountability, transparency, and the potential for ethical violations in warfare.
Social Isolation: Increased reliance on AI-powered communication and virtual interactions may lead to decreased face-to-face human interaction, potentially resulting in social isolation, reduced empathy, and deteriorating mental health.
Deepfakes and Misinformation: AI technologies can be used to create convincing deepfake videos or generate synthetic text, making it increasingly difficult to discern truth from falsehood. This poses a significant threat to trust in media, public discourse, and democratic processes.
Economic Disparity: While AI has the potential to create economic growth, its benefits may not be evenly distributed. Wealthier individuals and corporations may have greater access to AI technologies, exacerbating existing economic disparities and creating a digital divide.
Lack of Accountability: AI decision-making can sometimes lack transparency, making it challenging to hold AI systems and their developers accountable for their actions. Lack of transparency can erode public trust and hinder the ability to address potential biases or errors in AI algorithms.
It is crucial to recognise and address these potential harms proactively by implementing robust regulations, ethical guidelines, and responsible AI development practices to ensure that AI technologies are developed and deployed in a manner that maximises benefits while minimising harm to society.