It's been a busy week. Apple revealed their aritificial intelligence play, Elon Musk brings to an end a months-long legal battle with OpenAI and what's the deal with large language models not being women friendly?
In other news, the head of the Catholic Church, Pope Francis, stepped out of the Vatican City to address the Group of Seven nations and other invited countries in a first for the religion's top authority. What did he speak about, why AI of course.
Here's a look at what happened this week.
WWDC Highlight: Apple Intelligence
Apple's got a really bad habit of leaving both journalists and its loyal fanbase on tenterhooks when it comes to announcing new features. Thank god for Bloomberg's Mark Gurman and his incredible sources within the tech giant.
After an agonising wait, the company at its annual Worldwide Developer's Conference announced its plans for AI: Apple Intelligence. If you've not spotted it already:
Artificial Intelligence = AI.
Apple Intelligence = AI.
Coincidence? Unlikely, given Apple's track record for inventing and then reinventing features.
The new technology platform from the $3-trillion company looks like a game changer.
Apple Intelligence is a new spin on AI and machine learning, pointing the technology inward at users and their device ecosystems. The company's new private cloud compute seems to a new way to access LLMs while also maintaining user privacy.
We did a full breakdown here.
AI's Bias Towards Women
Generative AI models and LLMs are no strangers to bias. Just ask UNESCO, which released a report earlier this year that singled out Llama 2 and ChatGPT-2 for their bias against women.
NDTV Profit caught up with Wipro's Global Chief Privacy and AI Governance Officer Ivana Bartoletti to discuss topics, including bias, data privacy, responsible AI and developing the technology in the Indian context.
So, does bias come from? Bartoletti puts it simply: "Garbage in, garbage out."
She's referring to the data that genAI models are trained on and the quality of the datasets. But even that, according to her, has a fix. If AI is learning from our behaviour, surely, we can find ways around the problem.
"We can use AI to identify potential correlations and causations between different things in society to identify sources of biases that we were not aware of and that we can't really see with human eyes," said Bartoletti.
You can read about the full conversation here.
OpenAI (1) — Elon Musk (0)
Who had OpenAI and Elon Musk beefing on their bingo card for 2024?
The tech billionaire moved to dismiss his breach of contract lawsuit against OpenAI and its Chief Executive Officer Sam Altman earlier this week. It's unclear why.
Musk's lawsuit filed in February this year alleged that the company had moved away from its goal of developing AI "for the benefit of humanity", and was headed in a for-profit direction, hence the breach of contract.
OpenAI has rubbished the claims, arguing in a court filing that the courts should dismiss the billionaire's plea and that Musk's lawsuit was an attempt to further his own AI company, xAI which is developing Grok, a "free speech" genAI model.
Following the filing of the lawsuit, OpenAI published several mail exchanges between the billionaire and the company. The evidence suggests that Musk and OpenAI were both roughly aligned on the requirement for a for-profit venture, a deviation from the billionaire's lawsuit.
Musk wrote in an email to former OpenAI employee and founder Ilya Sutskever, "Even raising several hundred million won't be enough. This needs billions per year immediately or forget it."
It was all but likely that the two would be at loggerheads sooner rather than later. Musk left the company back in 2018 after the board didn't take too kindly to him suggesting that he take over the company.
However, parts of the SpaceX founder's lawsuit do have some merit as they're based on the founding principles of OpenAI, which started as a non-profit organisation back in 2015. Prior to Musk's departure the company worked mostly as non-profit, before switching gears in 2019.
That year, OpenAI announced the formation of its subsidiary OpenAI LP, the for-profit arm which has claimed to work under a "capped profit" structure. OpenAI LP essentially allows the entitiy to legally attract investment from venture capital funds as well as grants employees stake in the company.
It's confusing why Musk dropped his lawsuit against OpenAI. Just a day prior, the billionaire was ranting about banning Apple devices from his companies after Apple announced its partnership with the AI industry leader. Guess we'll find out eventually.
DuckDuckGo's AI Chat
This one is from last week, but it's interesting enough to be in this week's wrap.
Privacy-oriented search engine DuckDuckGo has launched a Beta version of its AI chat feature. The way it works is that it allows users to have anonymous conversations with AI models without recording any of their data. So how do they do it?
The company ensures that all user metadata, like IP and DNS addresses are scrubbed before sending user queries to the LLMs. Currently, DuckDuckGo supports Anthropic's Claude 3 Haiku, Meta's Llama 3-70B and OpenAI's ChapGPT 3.5 Turbo to name some of the more popular models.
Even if users decide to use personal information while posing queries to the LLMs, DuckDuckGo makes sure to present the data as coming from itself, as opposed to specific users. This helps prevent models from creating personal profiles of individuals based on the questions they ask.
Further, the company claims that they have agreements in place with all the model providers that "further limit how they can use data from anonymous chats". Notably, DuckDuckGo require model providers to delete all information received after a maximum of 30 days from when the conversation took place.
It's hard to believe the company's claim, especially at a time when privacy is constantly being eroded in a world driven by AI. However, the company had a stellar record of offering its users extra layers of privacy.
DuckDuckGo's AI Chat feels like a breath of fresh air when privacy is a concern.
With inputs from Tejas Kala.
Beyond Tomorrow is a weekly newsletter sent to your inbox every Saturday to give you a roundup of everything AI in the last week.
RECOMMENDED FOR YOU
U-Turn? Donald Trump Wants Elon Musk, His Companies To 'Thrive' In US
Jul 24, 2025
Tesla In India: Online Orders Begin! Here's How You Can Book Your Tesla Y Model—Check Priority Locations
Jul 22, 2025
Elon Musk Brings Tesla To India — But Who’s It Really For?
Jul 15, 2025
Elon Musk’s xAI To Launch Grok 4 On July 9 — Here’s What To Expect
Jul 08, 2025
POPULAR