- Apple is eager to jump on the AI hype train.
- It trains artificial intelligence models using proprietary data without permission.
- Artificial intelligence uses too much power, even if some of it is done "on the device."
Apple is so panicked about being left behind on the AI hype train that all of its ethics are thrown out the window.
There are many reasons to use Apple products. They're beautifully designed and built, work well together, and more. Apple is also a company you can trust when it comes to protecting your privacy, shifting operations to renewable energy, and generally having a strong ethical stance compared to other companies in Silicon Valley. When Apple announced its plans at WWDC this month, many of us were waiting for Apple to take a greener, more ethical approach to artificial intelligence. Boy, were we wrong.
“Consumers pay premium prices for Apple products for many reasons, chief among them privacy. Apple’s brand is built on protecting users’ private data, even at the risk of losing critical functionality and limiting interoperability. In contrast, OpenAI A lifeline via email, says Gemma Galdón-Clavell, Founder and CEO of Eticas AI and Advisor to the United Nations and European Union on Applied Ethics and Responsible Artificial Intelligence.
Apple has come a long way with its on-device AI processing, using the Neural Engine that has been in iPhones, iPads and Macs for years. They run AI models locally, which is more private, and because they're purpose-built to run trillions of AI-type operations per second, they use less power than trying to use a CPU to perform the same operations. Much faster.
So far, so good. But once your iPhone or Mac can't handle an AI task, it kicks it over to Apple's new AI servers. We don’t know the technical details of its power usage, but as a guide, AI servers use about ten times more power than regular cloud servers and require more water (for cooling). Then, if your iPhone thinks the answer provided by a third-party service (initially ChatGPT) will be better for you, it will send you to Open AI's energy-consuming, carbon-emitting servers.
“Microsoft has committed to becoming carbon negative by 2030,” Charles Cadieu, founder of climate technology and carbon removal company Spiritus, told Lifewire via email. “Similarly, Apple is developing renewable energy projects with the goal of achieving a net-zero carbon footprint across its entire supply chain and product lifecycle by 2030. While these are solid goals, the pace of AI development and increasing carbon emissions are making these goals more difficult. "For example, Microsoft's 2023 emissions increased 29.1% from the 2020 baseline, mainly due to building more data centers to meet artificial intelligence needs. "
Remember, this is the same Apple that showed off a nature-themed promotional video during last year's WWDC keynote. So far, it has sent a clear message about energy waste.
We might also wish Apple could train its own AI models instead of stealing media from everyone on the internet, without asking and without any way to opt out. but not. Like every other company that uses the internet to train AI models, Apple’s training data is made up of our words and pictures. This is from Apple's own documentation:
“We train our base model on licensed data, including data selected to enhance specific features, as well as publicly available data collected by our web crawler AppleBot .” [emphasis added].
The best perspective I've seen on this policy is Javascript developer Melanie Kat's quip on Mastodon. "I have never pirated anything in my life, I just used publicly available data from the internet," Kate wrote. "Kind of like when you see a product on the shelf at an Apple store, you can pick it up and walk out. It's public after all."
You can opt out and block AppleBot on your site, but it's too late for any model Apple has already created. This only makes sense if you run a website that you control and know how to opt out.
Amid the ongoing climate disaster, skyrocketing energy use, complete disregard for privacy, ownership and copyright of materials on the internet, and its partnership with OpenAI, Apple seems to have just given up on its ethics when it realized it was lagging behind in artificial intelligence. It begins with a focus on privacy and environmental ideals. This is standard operating procedure for all other public corporate entities, but Apple always seems to have different ideas.
Brian Prince, founder and CEO of Top AI Tools, an artificial intelligence education platform, said: “For decades, Apple has modeled an ethical business model when it comes to apps. […] Apple also has Broad Code of Ethics and Standards” Via Email Lifeline. "In the race to create better, more efficient, and more useful AI systems, some companies are ignoring the ethical standards set for the technology. Ideally, Apple should seek permission before scraping the web, and should strive to retain as much as possible information." Their processing takes place on-device whenever possible to maintain user security and privacy. "
This looks like a huge misstep that could irreparably damage Apple's hard-earned reputation as the only "good" tech giant.