AI-powered accessibility tools like Meta AI, Be My AI, and Access AI from Aira are significantly enhancing how visually impaired users interact with the world. Each of these tools has distinct approaches, features, and benefits. Below, we compare these solutions in detail, including Aira’s new AI initiatives that are shaping the future of accessible technology.

Meta AI

Meta AI is a broad, general-purpose AI assistant integrated into Meta’s platforms, such as Facebook, Instagram, and WhatsApp. It leverages advanced language models like Llama to offer generative AI capabilities, including text, image recognition, and chat-based assistance. Meta AI’s strength lies in its powerful generative features and widespread integration, which makes it suitable for a wide range of everyday tasks beyond just accessibility.

However, Meta AI is not specifically tailored to the needs of visually impaired users. It focuses on general interaction improvements, and while it offers high-level image descriptions, it lacks the accessibility-specific refinements that specialized tools provide. Meta is currently expanding its AI reach but faces regulatory delays in Europe due to privacy and data use concerns. As part of its commitment to responsible AI development, Meta AI allows users to control data usage and offers transparency about its data handling practices.

Be My AI

Be My AI is a feature within the Be My Eyes app that uses AI, powered by OpenAI’s GPT-4 Vision model, to provide detailed descriptions of images. This tool complements the live assistance offered by sighted volunteers, allowing users to access quick and descriptive feedback on visual content. Be My AI’s strength is in its conversational style, where users can ask follow-up questions to gain deeper context about what is being seen.

The focus of Be My AI is on providing accurate and responsive descriptions specifically for visually impaired users. It excels in making AI interactions feel personal and relevant, offering a straightforward, user-friendly experience tailored to individual needs. However, unlike Aira’s Access AI, Be My AI does not offer human verification, which can be a critical feature for ensuring high trust in certain situations.

Access AI from Aira

Access AI is part of Aira’s broader vision of integrating AI into its existing visual interpreting services. It allows users, known as Explorers, to capture or upload images and receive instant AI-generated descriptions. What sets Access AI apart is the optional human verification through Aira Verify, where a professional visual interpreter can review and confirm the AI’s responses. This combination of AI and human input ensures that the service remains highly accurate, secure, and reliable.

Access AI also includes features like multi-photo upload, verbosity controls, and chat history, which enhance user interaction and personalization. Additionally, Aira’s commitment to privacy means that no Access AI sessions are shared with third parties, safeguarding user data. Aira’s new Build AI initiative further advances its AI capabilities by allowing users to contribute to AI development in a secure and controlled manner. This program, available primarily in the US, collects real-world data to improve future AI features, enhancing Aira’s service without compromising user privacy. The access AI is free for the time being.

Each of these AI tools offers unique benefits, catering to different needs and preferences. Whether you’re looking for a general-purpose AI assistant like Meta AI, a visually impaired-focused tool like Be My AI, or a hybrid solution with human verification like Access AI, there’s a tool that can help enhance accessibility in your daily life.

My Experience with Gemini vs. ChatGPT: Why Google’s AI Didn’t Meet My Expectations

When Google announced Gemini, I was excited about the prospect of a new AI that promised to elevate productivity and simplify daily tasks. With advanced capabilities like managing Google Calendar events and understanding complex queries, I believed it would be a valuable addition to my workflow. However, after trying Gemini, my experience didn’t match the high expectations set by Google’s marketing.

High Hopes from the Google Event

Google’s event showcased Gemini as a cutting-edge AI assistant designed to handle everyday tasks more intelligently. They highlighted its ability to manage calendar events with simple voice commands, interact naturally, and offer better contextual understanding. It felt like the AI we had been waiting for—a smart assistant that could handle scheduling, reminders, and more with ease.

I quickly set up the AI and started using it for tasks like scheduling appointments, checking my calendar, and managing daily reminders. Unfortunately, the reality didn’t live up to the presentation.

Disappointing Performance with Calendar Integration

One of the most anticipated features of Gemini was its ability to manage Google Calendar events. The promise was that you could simply say, “Add an event to my calendar,” and the AI would take care of the rest. However, my experience was far from seamless:

1. **Unreliable Command Execution**: The AI frequently misinterpreted commands or failed to respond altogether. This was especially frustrating when I was trying to quickly add or adjust calendar events. Instead of simplifying my scheduling, Gemini often made it more cumbersome, requiring multiple attempts to get a simple task done【8†source】.

2. **Dependence on Google Assistant**: Gemini heavily relied on Google Assistant to perform calendar tasks, which often led to confusion and delays. This dependence made the experience feel disjointed, as it wasn’t always clear whether Gemini or Assistant was handling the request. The reliance on Assistant highlighted that Gemini was not a fully independent AI solution, detracting from the seamless integration promised by Google【9†source】.

3. **Contextual Understanding Issues**: One of the biggest letdowns was Gemini’s lack of contextual awareness. It struggled to maintain the flow of a conversation, especially when handling follow-up questions or complex scheduling scenarios. This shortcoming made it difficult to rely on Gemini for anything beyond basic, straightforward commands.

Coding Challenges: Gemini Falls Short

As someone who frequently works with code, I was particularly interested in seeing how Gemini could assist in coding tasks. However, I found that Gemini struggled significantly when it came to providing accurate and useful coding assistance:

– **Limited Code Understanding**: Gemini often generated incorrect or incomplete code snippets, which required extensive corrections. It didn’t handle complex coding scenarios well and frequently misunderstood the context of what I was trying to achieve.

– **Poor Debugging Assistance**: One of my main frustrations was Gemini’s inability to effectively help debug code. Unlike ChatGPT, which can provide detailed explanations and suggestions for fixing errors, Gemini’s responses were often too vague or off-target to be helpful.

– **Inconsistent Code Formatting**: Even basic code formatting suggestions were inconsistent, making it difficult to rely on Gemini for any serious coding assistance. This was a significant drawback, especially when compared to the more polished and reliable performance of ChatGPT in handling code-related queries.

Why ChatGPT Outshines Gemini

After struggling with Gemini, I returned to ChatGPT, which consistently outperformed Google’s AI. ChatGPT’s strengths became clear in contrast:

– **Consistent Contextual Awareness**: ChatGPT handles complex queries with ease, maintaining context throughout conversations without needing constant corrections. This ability makes it far more reliable for tasks that go beyond simple commands.

– **Seamless Integration and Performance**: Unlike Gemini, ChatGPT works consistently across platforms and devices. Whether I’m using it on my desktop, phone, or through various apps, the performance is smooth and responsive, with no awkward transitions or delays.

– **Superior Command and Code Handling**: ChatGPT executes commands accurately and quickly, whether I’m asking it to manage a schedule, answer questions, or provide detailed responses. It’s especially strong in coding tasks, offering reliable code generation, debugging assistance, and well-formatted snippets that save time and effort.

My experience with Gemini was a reminder that high expectations don’t always align with real-world performance. While Google’s AI shows potential, especially with continued updates, it still has a long way to go before matching the reliability, contextual understanding, and ease of use offered by ChatGPT.

If you’re exploring alternatives or have suggestions, I’d love to hear from you! Leave a comment on this post or join the iAccessibility community, where we discuss various accessibility tools and strategies, including AI-driven solutions. You can join us

Home

 Let’s collaborate and find the best ways to make these tools work for everyone!

For now, ChatGPT remains my go-to assistant, delivering consistent results and adapting seamlessly to my needs. As AI continues to evolve, I look forward to seeing how both Gemini and ChatGPT improve, but until then, ChatGPT is the tool that truly supports my day-to-day tasks with reliability and precision.

If you want to talk AI, please email me at taylor@techopolis.online

Apple is adding artificial intelligence in in iOS 18.1 for iPhone 15 Pro models and up, and any other devices that support M series processors. One of the features coming is the ability to have notifications summarized on the Lock Screen through the use of Apple Intelligence

How Does This Work?

Sometimes an app may show several notifications on your Lock Screen, and you will need to expand the group of notifications to see what everything is about. This can take time and effort to click through all of your notifications, or to scroll through them if you are low vision. Notification summaries will show a summary of your notifications at the top of the stack, and there will be a number next to the app icon to show how many items are in that stack.

Example

Here’s how this is nice. I have a Discord bot that states who has joined my Discord server in a text channel, This allows us to know when someone is online, so that we can all join the conversation. Apple Intelligence saw several notifications and summarized them by saying that Multiple people have joined General including @Person1, and @person2. This makes it much easier to understand these notifications without moving through each of the items in the notification stack.

Conclusion

I think Notification Summaries are going to be extremely helpful to iOS users, and I look forward to seeing what they do in the full release. We are starting to see some major improvements to iOS through AI, and it will only get better from here.

Do you use Apple Intelligence, or have you tried this feature? Let us know in the comments, or anywhere in the community.

AppleVis announced that it would be closing down on July 27th, 2024. The AppleVis team stated that the site would stay in read only mode until August 31, 2024, and would then shut down.

On August 28th, 2024 AppleVis announced that they would be acquired by Be My Eyes, and that the site would continue, and that the site would re-open on September 9th. This means that AppleVis, and all of its wonderful resources will continue to exist after the shutdown date.

It is unclear at this point as to how the website will change after it comes back, but I think it is very exciting that this resource will continue to exist for everyone to take advantage of. There has been a lot of content posted at AppleVis, and I think it is important that it remains online.

You can learn more about this by reading the press release from Be My Eyes.

Be My Eyes Aquires AppleVis to Secure Its Future and to Invest For Growth

Introduction to AI

Artificial Intelligence (AI) is a technology that mimics human intelligence to perform tasks like understanding language, recognizing objects, and making decisions. While AI might seem complex, it’s quickly becoming an integral part of everyday life, offering incredible tools for everyone, including those who are blind or visually impaired. AI can help you navigate the world, access information, and communicate more effectively.

How AI Can Help You

For blind users, AI can be a game-changer, making many daily tasks easier and more accessible. Here are some of the most impactful AI tools you can start using today:

ChatGPT: Your Conversational Assistant

What It Does: ChatGPT is an AI you can talk to like a friend or assistant. You can ask it questions, get help writing emails, or even have it explain complex topics in simple terms.

How to Use It: To get started, visit chat.openai.com in your web browser. You can type or speak your questions, and ChatGPT will respond with helpful information. Whether you need to draft a message, learn something new, or just chat, ChatGPT is there to assist you.

Free Plan and Pricing:

Free Plan: ChatGPT offers a free tier that gives you access to the basic features of the AI. This plan is great for getting started and exploring what the AI can do without any financial commitment.

ChatGPT Plus: For those who want more advanced features, faster response times, and priority access during high-demand periods, OpenAI offers a paid plan called ChatGPT Plus. This plan costs $20 per month and provides enhanced capabilities, making it ideal for users who rely on ChatGPT for frequent or complex tasks.

Getting Started:

You don’t need an account to start using the free version of ChatGPT, but signing up can provide a more personalized experience. If you’re new to AI, I highly recommend exploring the free plan first to see how it can fit into your daily routine. As you become more comfortable, you can decide if upgrading to ChatGPT Plus is right for you.

Seeing AI: Your Eyes for the Visual World

What It Does: Developed by Microsoft, Seeing AI is a free app that narrates the world around you. It can read text, describe people and objects, and even identify colors. For example, you can use it to read a letter, check the color of your clothes, or even recognize the faces of people in photos.

How to Use It: Download the Seeing AI app from the App Store. Once installed, you can use your phone’s camera to point at text or objects, and the app will describe what it sees. Seeing AI is particularly useful for handling tasks like reading grocery labels, identifying ingredients, and even recognizing currency.

Be My AI: A Companion for Visual Assistance

What It Does: Be My AI is a feature within the Be My Eyes app, which connects blind users with sighted volunteers via video call. With its integration of GPT-4, Be My AI can now provide instant descriptions of photos, identify objects, and help with tasks like setting up a new device or reading product labels—all through AI, without needing a human volunteer.

How to Use It: Start by downloading the Be My Eyes app. Within the app, you can use the Be My AI feature to ask the AI to describe what’s in front of your camera. This is incredibly useful for quick, real-time assistance with visual tasks.

Trust But Verify: A Note on AI Limitations

While AI can significantly enhance your daily life, it’s essential to remember that it isn’t perfect. AI tools like ChatGPT, Seeing AI, and Be My AI are incredibly advanced, but they can sometimes make mistakes or misinterpret information. Therefore, it’s always a good practice to trust but verify. Double-check important details when possible, and use AI as a helpful assistant rather than a sole source of truth.

Everyday Uses of AI

AI isn’t just for special occasions—it can be part of your daily routine:

• Morning Routine: Check the weather, get news updates, or listen to your schedule for the day using AI tools.

• Shopping: Use AI to read labels, compare products, or find the best deals.

• Traveling: Get AI-assisted navigation to ensure you reach your destination safely and efficiently.

• Entertainment: Discover new books, movies, or music recommendations based on your preferences.

Resources

• ChatGPT: https://chat.openai.com

• Seeing AI: https://apps.apple.com/us/app/seeing-ai/id999062298

• Be My AI: https://www.bemyeyes.com

Did you know you can type tye  (Apple) logo on any Apple device with a keyboard? Here’s how to use it.

1. Enter a text field on any Apple device that has a keyboard.
2. Use the keystroke Option + Shift + K and you will see the  logo.

NOTE: This may not work on other platforms including Windows or Android.

This podcast episode from RNIB features a demonstration of the current version of the iAccessibility app.

RNIB Tech Talk iAccessibility Demo Episode

Listen to this segment on Kelly and Ramya’s show on AMI, where Michael Babcock interviews Michael Doise about the current roadmap of iAccessibility.

Kelly and Ramya – We learn about iAccessibility

There has been a lot of changes that have happened here at iAccessibility this month, and there will be many more to come. In this post, I want to discuss where we’ve been, and where we’re going as iAccessibility so that we are all on the same page.

The Past of iAccessibility.

iAccessibility started in 2010 as a blog to review apps called the iAccessibility Report. I, Michael Doise, reviewed apps with VoiceOver on this blog to determine what worked with VoiceOver. The blog then started to have other content regarding technology, and we started a podcast in 2015 called the iACast. In 2016, we started our WhatsApp Community and TeamTalk Server, which later became a Discord community server instead. iAccessibility has also had the chance to work with some amazing podcasters like Unmute Presents and many others.

Where We Are Going

iAccessibility has been a part of my life for the past 14 years, and the goal has always been to build a thriving community of technology users who help each other in the efforts of learning and using technology. The below items describe where we are, and where iAccessibility is going.

• iAccessibility has built an app directory so that blind and low vision technology users can find out if an app is accessible, or at least usable. This app directory covers Apple platforms, but also includes Windows and Android devices as well. We will have a similar layout to what may expect, but we will also have features like AI based search and more.
• The iACast will continue, and we will be improving it with more conversations, and interactive experiences. These may include town halls, and interactive calls on Zoom and Clubhouse
• iAccessibility has created a new forums system using Discourse. We plan to work with Discourse to increase the accessibility of these forums, so that other forum communities can benefit from these accessibility changes.
• Bug Tracking is very important, and we plan to start tracking bugs with Apple platforms. We are considering a plan to create a bug tracker with BugZilla for users to submit bugs through, and these will be reflected through a blog post on the iAccessibility website.
• We will also be opening up opportunities for users to contribute. Right now, users can contribute to the App Directory, but we will soon allow users to submit posts to the Report blog. or to submit podcast content. All posts will need to be approved, but this allows the site to be ran and driven by the community.
• iAccessibility can’t be ran by one person and one person should not run a site as big as iAccessibility. Therefore, the team has decided to create a nonprofit organization ro run and manage iAccessibility. The board has been named, and documentation will be filed to form the nonprofit in the coming weeks. This will allow the team and community to make decisions to drive the forward progress of the website, forums, and other services provided by iAccessibility.

Conclusion

iAccessibility has always been a place to help people to connect with each other, and learn to use technology. It is my hope that this roadmap will help everyone know where we’ve been, and to help you all know where we are going in the future. We have been around for 14 years, and I think the next 14 will be even more amazing with all of you building what iAccessibility is going to be. Thank you for reading this, and please feel free to reach out and ask us any questions you have have either on Mastodon at https://iaccessibility.social/@iaccessibility, on X as @iaccessibility1 We are still working on emails, but you can email mikedoise@icloud.com until everything is set up. Thank you again, and I look forward to speaking with everyone in the community.