Another week, another controversy. If nothing else, OpenAI is in the news pretty often, and that's without its biggest investor making announcements of AI-powered computers. Plus, how does one solve the problems that crop up when building AI in a country like India. Have a look.
Controversy courts OpenAI
For a company that started out as a non-profit researching and creating artificial intelligence, OpenAI has found itself mired in controversy over the past two weeks.
Earlier this week, actress Scarlett Johansson released a statement through her agency saying that she was contacted by OpenAI founder Sam Altman himself in September last year about voicing the audio feature for their chatbots, Sky. Johansson declined.
Despite that, last week's GPT-4o announcement showcased a demo of the Voice Mode, which sounded eerily like the actress' AI character from the 2015 science fiction film, Her.
Altman has declined that they used Johansson's voice and said in a statement that the voice was "never intended to resemble hers." The company has since taken down the voice called Sky and replaced it with another one called Juniper. In a blog post on May 19, the company said that the voices had been picked much before Altman got in touch with the actress and that Sky's voice belonged to a different professional actress who was using their "natural speaking voice."
In other news, there's been another high profile exit from OpenAI. This time, its one of their policy researchers Gretchen Krueger. In a thread on X (formerly known as Twitter) the AI policy researcher highlighted concerns she had about OpenAI's accountability, transparency and documentation practices.
For a company that has been on the cutting edge of creating generative AI, the company continues to have reality television levels of twists and turns. It's likely that all of this, barely six months after Altman's ousting incident, will have investors eyeing the company with a little bit of caution.
I gave my notice to OpenAI on May 14th. I admire and adore my teammates, feel the stakes of the work I am stepping away from, and my manager @Miles_Brundage has given me mentorship and opportunities of a lifetime here. This was not an easy decision to make.
May 22, 2024Microsoft's AI laptops
Microsoft has introduced a new category of AI-first Windows computers called Copilot + PCs. According to the tech giant, these systems are the "fastest, most intelligent Windows PCs ever built."
The company is bringing the technology to Microsoft Surface systems and has partnered with the likes of Lenovo, Acer, Dell, HP, ASUS and Samsung to bring the tech to consumers starting at a price of $999 (~ Rs 83,000).
The new systems will enable Copilot+PCs to run AI workloads at least 20 times faster and 100x more efficiently than the usual PC, according to the Microsoft press release. The reason behind the speed can be attributed to the fact that the company is adding a neural processing unit to the incoming computers. These chips will help the computers connect with the large language models running in Microsoft's cloud computing platform, Azure.
What makes it, depending on your perspective, cooler or spookier, is that the AI experience will come with a new feature called Recall. How does it work? Anything you've ever seen or done on your computer, Copilot+PCs will be able to pull that up for you through a simple search. Here's a quick look at how it works.
Do you want a photographic memory? The next best thing is Recall. Check out this new experience only available on Copilot+ PCs. Here is a demo from my colleague Carolina. pic.twitter.com/goylSXxslH
May 20, 2024Awesome or terrifying? Depends on who you're asking and how you feel about your privacy. Microsoft says that the data Recall uses to pull up content from your computer browsing is all processed and stored locally. It also gives you options to turn off how and what Recall remembers so you have full control.
AI wearables company looking for buyer
AI startup Humane is looking for someone to buy its business, Bloomberg reported. The news comes just weeks after the company's wearable AI device had a less than stellar launch.
Common complaints about the company's AI Pin, which you wear like a accessory on your clothes include its battery life, the device overheating, accuracy and response time. Its hefty price tag of $699 hasn't helped with its lukewarm reception either.
Started in 2015 by two longtime Apple veterans, the company has managed to raise $230 million in funding from several high profile investors, including OpenAI's Sam Altman. The company was valued at $850 million, according to The Information.
Bloomberg reports that the company is currently seeking anything between $750 million and $1 billion in the sale and is currently working with a financial advisor on the next steps forward.
Indian IT must brace for Gen AI
Disruption has been a long time coming for India's leading tech companies. With the way AI and generative AI is being talked about and adopted, they might have to either adapt or make way for smaller, more agile companies.
NDTV caught up with ChrysCapital's Sanjay Jalona and Persistent Systems' Sandeep Kalra to figure out how they feel about Indian IT's ability to deal with the incoming "AI Wave".
Jalona breaks down three reasons why he believes Indian IT won't have that problem:
1) The ecosystem is ready for GenAI adoption due to prior investments in chipsets, datasets, and language models.
2) How deeply GenAI has already penetrated into our lives, from children using it for homework to adults planning the cheapest vacations possible.
3) Businesses have identified that GenAI has the potential to boost operational efficiencies and are driving investments into the technology to stay ahead in the game.
You can read about the full conversation here.
Building AI in India
In a country with 121 languages and 270 mother tongues, building a large language model is more than likely to become an arduous affair. In fact, one of the bigger complaints made by Indian tech companies building generative AI in the country has been the fact that data sets simply aren't available. Accounting for India's linguistic and cultural diversity are only the tip of the iceberg when it comes to solving the issues around building AI.
Currently, there are only two companies in India that are working on creating such systems from the ground-up: BharatGPT and Krutrim.
“We're lacking a lot in terms of diversity of datasets in India, specific to India, collecting data faster, improving processes to collect it is a challenge,” said Akbar Mohammed, head of Fractal Dimension, which is an interdisciplinary team that works to solve problems regarding things like responsible AI, sustainability, Gen AI adoption strategy and more.
You can read the full story about the challenges involved in building AI for the Indian context here.
Beyond Tomorrow is a weekly newsletter sent to your inbox every Saturday to give you a roundup of everything AI in the last week.
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.
