Free General Purpose Generative Artificial Intelligence Tools
As the legal profession continues to evolve and adapt to recent technologies, it’s becoming increasingly important for lawyers to understand and leverage the capabilities of generative artificial intelligence (AI) tools. These tools, which can assist with tasks such as data analysis, research, translation, image and video creation, summarization, transcription, and automation, have the potential to increase efficiency and productivity within the legal field.
To get a sense of how to leverage these tools without a monthly subscription here are a few of the main models that are available for free. Getting used to writing prompts and figuring out what these tools are good at will help you understand how they will eventually help you in practice. While it’s crucial to be aware of the ethical implications, privacy impact, and potential biases associated with these tools, learning to effectively utilize them can provide a significant advantage in practice.
Ethical Considerations
While generative AI tools offer many potential benefits for lawyers, it’s crucial to approach their use with caution and ethical consideration. Lawyers should always fact-check and verify the information provided by these tools, as they can occasionally produce inaccurate or biased outputs. Additionally, it’s important to ensure that no confidential or sensitive client information is inadvertently shared or exposed when using these tools. Generative AI should be viewed as a powerful assistive technology, but not a substitute for a lawyer’s professional judgment and ability.
ChatGPT-4o
From OpenAI, the latest ChatGPT is the ChatGPT-4o (Omni), a multimodal model which can handle audio, text, and images. Multimodal models, like ChatGPT-4o, are capable of processing and generating outputs across multiple data formats, such as text, audio, images, and video. This versatility makes them particularly useful for a wide range of tasks and applications. It offers GPT-4 level performance and is the first time ChatGPT users can use the GPT-4 model for free.
To use ChatGPT-40 you will need to create an account or use your existing account and go to https://chatgpt.com/?model=gpt-4o . People who pay for ChatGPT-4o will get an enhanced version, but there is a very capable free version. The underlying training data set was last updated in October of 2023, but you can also use it to browse the web. Here are some examples of what you can do, from data analysis to research to image generation to text generation.
There are certainly privacy implications, and with the anticipated integrations into iOS18 you will need to understand how to limit the data collection and app permissions. In your ChatGPT account go to Settings and there you can manage archived chats, delete chats, turn off “Improve the model for everyone” in the data controls and add multi-factor authentication.
Claude3
Claude is the generative AI tool built by Anthropic. Like ChatGPT, there are several models, with the Claude3 Sonnet providing free access to everyone. Claude3 is multimodal, capable of processing text and image data, and can transcribe and analyze static images including handwritten notes. Claude3’s data set was trained on data up until August 2023, so it will be unaware of information after that date. You can upload up to 5 documents or images, 10MB each. Claude 3 outperforms ChatGPT in benchmarking tests for reading complex PDFs.
You will need to create a login to use Claude. Unlike ChatGPT, you have no control in the free version to opt out of saving your prompts and user history, other than deleting your account.
Google Gemini
In December of 2023 Google released the update to their initial salvo into generative AI (Bard) with the product from their DeepMind project called Gemini. There are many enhancements including the ability to reason in Google Workspace products like Drive and Gmail, as well as the release of the Google NotebookLM, a “smart” notebook. Gemini can “reason” across text, images, video, audio, and code. It can run on everything from data centers to mobile devices.
To access Gemini, go to https://gemini.google.com/app. You will need to sign in with a Google account. By default, Gemini saves your activity, and trained reviewers process your conversations to improve the model. However, you can turn off Gemini Activity at https://myactivity.google.com/product/gemini?pli=1 to prevent your conversations from being reviewed and used for model improvement. It’s important to note that if you choose to turn off Gemini Activity, Extensions will also be disabled.
If you are a heavy user of Google products, one of the more interesting and exciting aspects of the Gemini update is the addition of extensions. For users of free Google products you can toggle on extensions for Google Flights, Hotels, Maps, Workspace (including Gmail, Drive, and Docs) or YouTube. Toggle on the extensions you want to use, and you can add an @ to mention a specific extension. For instance, “tell me the decision the HOA Board came to regarding use of windows units @gmail”. You can turn Gemini Apps Activity off if you don’t want future conversations reviewed or used to improve machine-learning models by going to Gemini Apps Activity . If you choose to turn off Gemini Activity, Extensions are turned off. Read the entire privacy question/answer page for Extensions to understand what you are and are not sharing. The privacy policy states that personal content in Google Workspace is not accessed or reviewed by human reviewers, used to train the model, or stored.
If you use a paid tier of Google products there are AI add-ons. In Google Workspace you can get Gemini for Workspace for $20 additional dollars per user per month for a one year commitment. You can also add Gemini Advanced for Google One subscriptions for $20 per month.
Microsoft Copilot
Copilot , the Microsoft generative AI, comes in several different forms, like the Google Gemini options. A free version of Copilot is available through the Edge Browser (f/k/a Bing Chat), or you can visit the website, and there is a free Copilot for Windows. Also, like Gemini, you will need to login to a Microsoft account. If you use a free account, your terms of service are different from the Microsoft 365 Business subscribers. Pricing is a bit confusing. Of note, there is Copilot for individuals and Copilot for Microsoft 365. Copilot for Microsoft 365 is an additional $30 per user per month for business subscribers and $20 per user per month for personal subscribers.
Copilot is multimodal, meaning that it can receive text, audio, and image inputs and output text, audio, and images. Of note, both Google Gemini and Microsoft Copilot’s AI tools provide links to the responses if generated from a web search in addition to their dataset. Depending on the prompt the links may be embedded or appear as a bibliography.
For Microsoft 365 subscribers logged into Edge, you can prompt from text that is displayed on the page, including content in Microsoft Word, Excel, PowerPoint, OneDrive, and Outlook if it is opened in the browser.
Perplexity
Perplexity bills itself as an alternative to traditional search engines by being a “conversational answer engine” so that you can get a concise answer with a curated set of resources. In addition an answering questions, creating code, and generating emails, you can interact with your data by uploading files, organize threads in collaborative collections by project or topic, and the responses to queries always cite their sources.
The underlying data source is a combination of the GPT dataset and Bing search engine. Perplexity Pro gives you access to OpenAI models (GPT-4 Turbo and GPT-4o) and Anthropic models (Claude3 Sonnet and Opus) for $20 per month or $200 a year with unlimited search, unlimited file uploads, and access to the upgraded AI models.
To use Perplexity, you need to create a login. As always, read the Terms of Service. There is a desktop app, as well as iPhone and Android apps. The Search Companion is a Chrome Extension and can be used like the Microsoft Edge and Gemini for Chrome extensions.
Once logged in go to Settings to choose whether to allow Perplexity to use your searches to improve the AI models. You can delete your account. You can also create a profile to get more personalized answers.
On the home page you can add your prompt. Click “focus” below the search box to choose whether you want to search the entire internet, only published academic papers, Wolfram|Alfa, YouTube, Reddit, or focus on writing without searching the web. You can also attach a limited number of images, text, or PDFs.
Perplexity combines traditional search tools with large language models like GPT-3. It links to sources and answers questions. For instance, here is a brief response to a simple prompt.
Poe
Now that you have a familiarity with some of the different large language models, if you want to do a side-by-side comparison of results then try Poe.com. You can prompt various versions of ChatGPT, Claude 3, DALLE 3, Gemini, Stable Diffusion, Llama by Meta, and many more. If you subscribe to Poe ($16.67 per month) for the annual plan, you get access to GPT-4 AND Claude 3 Opus (each a $20 subscription). Poe’s parent company is Quora.
On the home screen create a prompt. You can add attachments, and give it voice commands. Then choose which LLM you want to use. At the bottom of the results, you can then compare the response with other LLMs so that you do not have to switch between different platforms.
You can use Poe to create bots. Choose your base bot (ChatGPT, Gemini, etc.), create a bot server, prompt the bot on how to behave and respond to user messages, add an alternative knowledge source (files or text) and finish the profile to create your own bot. It is free and an interesting exercise to go through as the future of how people will interact with information (see example: Lowenstein Sandler Launches Custom GPT Chatbot | Lowenstein Sandler LLP).
Conclusion
Generative AI is rapidly becoming integrated into many of the technologies lawyers rely on daily. By learning how to effectively use these tools, create clear and specific prompts, and understand their underlying capabilities and limitations, lawyers can use generative AI to streamline workflows, enhance productivity, and gain valuable insights. However, it’s essential to approach these tools with a critical eye, prioritizing ethical considerations and maintaining a steadfast commitment to protecting client confidentiality and providing accurate, well-reasoned legal advice and analysis. What can you do with generative AI? Attorney Mitch Jackson lists 1000 ways, you just have to come up with a few to get started.