ChatGPT session for IOMP-IUPESM-ISC

Date

August 29, 2023

Since the launch of ChatGPT in November 2022, the hype around generative AI (artificial intelligence) has been immense. Products that can automatically generate text, images and video when prompted to do so are being discussed as a game-changer for organizations and society at large.

Generative AI promises organizations and their staff a route to increase their productivity, supporting them in automating – or at least delivering more quickly and easily – many important regular tasks.

This practical workshop aimed to demonstrate ways to experiment and explore how the most popular tool – ChatGPT – can be used in everyday work. It offered insights on how to work around technical capabilities and limitations of the tool, what makes for an effective prompt, and answer practical questions around the use of ChatGPT.

Session take-aways

Agenda and prompts

‣

1. Key definitions

❓

AI - Artificial Intelligence The development of computer systems that can perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and language translation.

❓

LLMs - Large Language Models A type of machine learning model used in generative AI that is trained on massive amounts of text data to generate new and original content.

❓

Generative AI A subtype of AI that involves using machine learning models to generate new and original content, such as text, images, or music.

❓

GPT - Generative Pre-trained Transformer A type of LLM developed by OpenAI that is widely used in generative AI applications. The basis of ChatGPT.

‣

2. Deep dive: ChatGPT

‣

Tips and hints for prompting

1️⃣

Roles Tell ChatGPT what role to play based on your needs. For example, you can ask ChatGPT to be a co-designer, an opportunity engine, a coach, a critical friend, a personal trainer, an impersonator, a futurist, or a risk assessor.

2️⃣

Context Provide detailed background information and content to help ChatGPT understand the situation and generate relevant responses.

3️⃣

Converse Have a conversation with ChatGPT and ask for changes, improvements, or clarifications on the generated content.

4️⃣

Analyse Identify strong and weak points in the generated responses and provide feedback to ChatGPT to help it improve.

5️⃣

Focus Use ChatGPT to support your areas of expertise and generate content that aligns with your goals.

6️⃣

Polite Be engaging and friendly with ChatGPT to establish a positive working relationship.

7️⃣

Guide Provide specific information and examples of outputs you want to help ChatGPT understand your needs and generate more relevant content.

8️⃣

Separate Use different chats for different topics and name each chat to keep track of the generated content.

‣

Limitations

1️⃣

2021 data cut-off ChatGPT is trained on text data up until 2021, so it may not have knowledge of recent events or developments.

2️⃣

No internet access ChatGPT does not have access to the web, unless you are using Browsing mode in the pro version.

3️⃣

Eager to please and makes stuff up ChatGPT may generate content that is inaccurate or misleading in an effort to please the user.

4️⃣

Different GPT models There are different versions of GPT models, such as GPT3.5 and GPT4, with varying levels of intelligence and features.

5️⃣

You won’t get the same answer twice ChatGPT generates original content each time, so you may not get the same answer twice.

6️⃣

Free version is often inaccessible The free version of ChatGPT may be inaccessible due to high demand or other factors.

7️⃣

Limited memory (context window) ChatGPT can only consider a certain amount of context when generating content.

‣

3. Try-it-at-home example uses of ChatGPT (and some other tools)

‣

ChatGPT as a Assistant

Summarise the transcript of a conversation (CLAUDE)

Act as: A minute taker

Claude.AI

Fill a cell with a maximum number taken from a range of cells

Act as: An Excel formula wizard

Daily reflection exercises to celebrate victories, learn from challenges, and stay connected with your purpose

Act as: A personal motivator

SPOKEN: Rewrite an article in very complicated Academic language into plain English. Then translate it to Spanish

Act as: An editing translator

‣

ChatGPT as a Partner

Evaluating the potential risks and ethical considerations associated with a proposed research study

Act as: A risk assessor

Scholarly

Scholar Assist

Scholar AI

Identifying case studies for a complex scientific problem in one field from another

Act as: A lateral search engine

Scholar AI

Scholarly

Understanding future scenarios for scientific publishing to plan potential strategic responses

Act as: A futurist

Bing AI

Emulating a peer reviewer's perspective to provide critique on a research paper

Act as: A persona / impersonator

Providing a guided learning experience on building my digital skills

Act as: A professional tutor

Testing and strengthening arguments about the impact of generative AI on science communication

Act as: A critical friend

Assisting in the formulation of a research proposal, including hypothesis generation and experiment design

Act as: A co-designer

Scholar AI

Scholarly

Generate alternative research methods or techniques to test a hypothesis

Act as: An opportunity engine

Scholarly

Scholar AI

Improve your negotiation skills when asking your manager for a salary raise

Act as: A negotiation coach

Be an English tutor for a Spanish speaker

Act as: A language teacher

‣

ChatGPT as a Team member

Create a personal marketing launch plan for an academic article

Act as: A content strategist

Analyse the registrations at a webinar

Act as: A data analyst

ChatGPT Advanced Data Analysis

Analyse a dataset on life expectancy over time

Act as: A data analyst

ChatGPT Advanced Data Analysis

‣

4. Risks and concerns around Generative AI

Misuse for misinformation/disinformation - Generative AI can be misused to spread false or misleading information.
Bias in outputted information - Generative AI can perpetuate biases present in the training data.
Data privacy - The inputted data used to train generative AI models may be sensitive and require protection.
Credibility and reputational risk - Misuse of generative AI can damage the credibility and reputation of individuals or organizations.
Environmental impact of training and using LLMs - Training and using LLMs requires significant computing power and may have environmental impacts.
Copyright of training data and produced outputs - Ownership and licensing of training data and produced outputs can be complex and require legal considerations.
Reputational and credibility risks if misused - Misuse of generative AI can result in reputational and credibility risks for individuals or organizations.

‣

5. Generative AI Tools database

AI Meeting Transcription

AI chatbot

Need to use Edge Browser and have a Microsoft account

Bard

Not available in EU

Claude.AI

Currently only available in the UK and US.

AI graphics

Microsoft Designer

Canva

Adobe Firefly

AI image creation or manipulation

DALL-E

Midjourney

Used within the Discord app

Ideogram

AI presentations

Gamma

Beautiful.ai

AI text to speech (TTS)

Automation

ChatGPT Plugins

ChatGPT Advanced Data Analysis

Included in ChatGPT Pro subscriptions, used to be called “ChatGPT Code Interpreter”

Copywriting

Document chat

6. Some suggested next steps

Test some of the examples provided in this session and explore the capabilities of generative AI.
Follow some AI influencers and stay up-to-date with the latest developments in the field.
Identify users within your teams who could benefit from using generative AI tools and explore potential use cases.
Develop safe spaces for experimentation in your team(s) and encourage innovation and creativity.

7. Downloads

Gamma.ai Presentation - key take aways (created during the workshop)14904.4KB

Contact Nick Scott

LinkedIn: linkedin.com/in/nicholasmscott

Email: nick@nickscott.digital