AI Adoption, On-device AI, Battle of Tech Giants, and Meta’s Backlash

Welcome back to yet another issue of Blindspot AI’s newsletter covering some of the biggest developments in the AI sphere.?

Whether you’re a business owner or a tech enthusiast, this issue is for you!

Estimated time of read: 5 minutes

AI adoption is picking up major momentum

McKinsey recently released 2 reports mapping the AI adoption trends. A must-read for any business leader or tech enthusiast:

• “Tech services and generative AI: Plotting the necessary reinvention”?
• “The state of AI in early 2024: Gen AI adoption spikes and starts to generate value”.

Here are key highlights the reports emphasize:

• The emerging market for services relating to gen AI/AI could be worth more than $200 billion by 2029 (Gartner actually forecasts AI Service spendings at $443 billion by 2027!)
• 65% of companies use Generative AI in their operations. Businesses report cost reductions and revenue increases, especially in marketing, sales, and product development.?
• 72% of companies have already adopted AI in some way.
• By active adoption of AI, companies can raise their profits up to 30%
• Focus on data-driven and, scalable AI solutions for your organization?

To stay competitive, business leaders need to consider investing in AI, acquiring specialized talent, and forming strategic partnerships to succeed in the new landscape.

How far are businesses with GenAI adoption??

According to McKinsey, they fall into three stages:

1. Observers (50-60% of businesses): Focusing on AI readiness and small-scale POCs for productivity.
2. Front-runners (30-40% of businesses): Implementing AI to reduce costs at scale, with AI centers of excellence and specialized roles like chief AI officer.
3. Innovators (<10% of businesses): Using gen AI for cutting-edge applications, driving significant business growth, and implementing an integrated AI strategy.

High-performing organizations are leveraging GenAI across multiple functions, focusing on risk management, data governance, and agile operations to maximize AI's potential.

All this means that by adopting AI at scale and strategically, businesses can now gain a significant advantage over more than 50% of the market that is still experimenting with small-scale POCs.?

If you want to find out more about the next steps for AI adoption in your business, contact us at [email protected].

Generative AI is now On-Device

In the past month and a half, the tech giants seem to have made their intention with Generative AI clear: “Bring it to the users sooner rather than later!”.

One after another “the big 3” (Google, Microsoft, and Apple) announced integrating complex AI solutions (utilizing their respective or partnered AI models) straight onto their devices. Either right into the operating system, as a premium feature, or as a built-in app.

Apple Intelligence for IoS, Copilot for Windows, and Gemini Nano for Android. LLMs are starting to appear on-device as a competitive edge.

This step brings AI even closer to users by integrating it into the “everyday use” of our smart devices like phones and PCs. Whether this is a good thing or not is a debate for another day, but it’s certainly an important thing to note whether you are a business owner or an everyday user.?

Apple’s Annual Worldwide Developers Conference: Apple Intelligence and Apple cloud?

Ever since the Generative AI boom started with the release of OpenAI’s ChatGPT, Apple seemed to lag behind its competitors, taking on a “more careful” approach when it came to AI.

However, at Apple’s annual Worldwide Developers Conference 2024 (available to watch here), the iPhone manufacturer shocked everyone by announcing the release of Apple Intelligence, a suite of AI features integrated straight into iPhones, iPads, and Macs.

Apple x Open AI Partnership?

Apple Intelligence integrates multiple generative models that have been fine-tuned to adapt to users' current activity and enhance their experiences. A significant component of this is OpenAI’s ChatGPT.?

The partnership between Apple and OpenAI made a notable impact on the tech industry as this collaboration raises questions about its implications for major players like Microsoft, Google, and smaller AI firms.

Apple Intelligence (tapping into ChatGPT’s capabilities) aims to improve user experience with:

• Writing and refining text
• Prioritizing and summarizing notifications
• Creating images
• Taking in-app actions to simplify interactions across apps
• and more…

One could say, Apple jumped right into the deep waters with this one, as it is arguably the first attempt to integrate third-party Generative AI into a device of this magnitude.

Security and Apple cloud

To keep user data safe, Apple created Private Cloud Compute (PCC), a cloud intelligence system designed specifically for private AI processing. (This was another unexpected step, as Apple has long championed on-device processing as the cornerstone for the security and privacy of user data.)

PCC is designed to keep personal user data safe by not being accessible to anyone other than the user (not even to Apple).

EU Regulation

European markets might not be seeing Apple Intelligence anytime soon due to potential conflicts with the EU Digital Markets Act (DMA).

The Act puts strict requirements on the “gatekeeper” companies that control massive online platforms to block anticompetitive behavior. The rollout delay might affect businesses relying on Apple’s ecosystem for their operations in Europe.

Annual Microsoft Build Conference: Copilot+ PCs

At the annual Microsoft Build conference (available to watch here), the tech giant introduced Copilot+ PCs, PCs with their AI assistant (Copilot) integrated directly into some Windows computers.?

Boosting their own AI product in the process, Copilot is designed to leverage advanced AI to help users manage their tasks, automate workflows, and provide insights into data.?

However, whether this will prove useful or instead turn into a security liability for users is still a question.

Google: Gemini Nano

To conclude the triad, in May 2024, Google launched Gemini Nano (on-device), a “cutting-edge” AI designed to enhance user interactions on Android devices.?

This AI is designed to provide context-aware assistance, helping users with everything from daily tasks to complex problem-solving, and to make AI more accessible and useful for everyday applications.

(So in theory, similar to Apple Intelligence with the difference of execution, as this is powered by Google’s slightly controversial LLM model, Gemini.)

Meta Starts Training AI on User Data. EU Pushes Back.?

Meta decided to use personal user data like pictures and posts from their Facebook and Instagram accounts to train its large language AI models.

However, for EU residents, the company faced legal pushback, forcing it to pause its AI model’s training launch in Europe.?

Meta also allows users to opt-out of their data being used to train AI models. Here is a tutorial on how to opt-out, if you do not wish your photos and posts to be used as AI training material.

Meta Watermarking AI Speech

Meta has found a way to watermark AI-generated speech. The tool is called AudioSeal and is available for free on GitHub. It can embed hidden watermarks in AI-generated speech to help detect and identify such content.?

This innovation aims to combat the misuse of voice cloning for scams and misinformation by allowing pinpointing of AI-generated segments within audio files.?

Despite its high detection accuracy, widespread adoption faces challenges, including the ease of tampering and the need for voluntary watermark application by creators.

That wraps up another edition of AI News. Stay tuned for the next issue and enjoy a great summer in the meantime!

Until next time,

Blindspot AI